Want to quickly access local running AI through Emacs and Ollama? This package provides a friendly Emacs interface for interacting with Ollama models and offers a convenient way to integrate to Ollama’s local LLM capabilities directly into your Emacs workflow with little or no configuration required!
“The name is just something a little bit fun and it seems to always remind me of the “bathroom buddy” from the film Gremlins (although hopefully this will work better than that seemed to!)”
I have a youtube channel where where I am looking to regularly post videos showcasing the capabilities of the package. Check it out here:
https://www.youtube.com/@OllamaBuddyforEmacs
See docs/ollama-buddy.org for the manual!
You can install with literally no config, but I would suggest binding to both menus.
(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper))A simple text information screen will be presented on the first opening of the chat (by selecting “o” from any of the menus), its just a bit of fun, but I wanted a quick start tutorial/assistant type of feel, here is what will be presented:
#+TITLE: Ollama Buddy Chat * Welcome to OLLAMA BUDDY #+begin_example ___ _ _ n _ n ___ _ _ _ _ | | | |__._|o(Y)o|__._| . |_ _ _| |_| | | | | | | | | . | | . | . | | | . | . |__ | |___|_|_|__/_|_|_|_|__/_|___|___|___|___|___| #+end_example (a) *tinyllama:latest* Select Info Pull Copy Delete (b) *qwen2.5-coder:7b* Select Info Pull Copy Delete (c) *deepseek-r1:7b* Select Info Pull Copy Delete (d) *codellama:7b* Select Info Pull Copy Delete (e) *gemma3:4b* Select Info Pull Copy Delete (f) *phi3:3.8b* Select Info Pull Copy Delete (g) *llama3.2:1b* Select Info Pull Copy Delete starcoder2:3b mistral:7b deepseek-r1:1.5b gemma3:1b qwen2.5-coder:3b llama3.2:3b - Ask me anything! C-c C-c - Main transient menu C-c O - Cancel request C-c C-k - Change model C-c m - Browse prompt history M-p/n/r - Browse ollama-buddy manual C-c ? - Advanced interface (show all tips) C-c A * *tinyllama:latest >> PROMPT:* <put prompt in here>
You will, of course, need to have ollama serve running, but technically, that is all you need! Emacs and ollama-buddy will handle the rest!
The example above has some models already pulled, but if you don’t have any initially, the Recommended Models section will be fully populated with suggestions. Just click, or move your point over, and press return, and ollama-buddy via Ollama will pull that model for you asynchronously!
Submit some queries, with different models:
PROMPT>> why is the sky blue? PROMPT>> why is the grass green? PROMPT>> why is emacs so great?
and show the token usage and rate being displayed in the chat buffer.
Open the Token Usage stats from the menu
Open the Token Usage graph
Toggle on ollama-buddy-toggle-token-display
See Screenshots / Demos section for more videos/demonstration.
Added Mistral Codestral support
- Added ollama-buddy-codestral.el for Mistral Codestral API support
- Implemented model handler registration for seamless switching between local Ollama and cloud-based Codestral models
- Supports async streaming responses with real-time token counting
- Follows the same integration pattern as other remote AI providers (OpenAI, Claude, Gemini, Grok)
- Uses “s:” prefix to identify Codestral models in the model list
(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:custom
(ollama-buddy-codestral-api-key
(auth-source-pick-first-password :host "ollama-buddy-codestral" :user "apikey"))
:config
(require 'ollama-buddy-codestral nil t))Simply bumped to version 1
Refactored content register processing to be more efficient and added a new Emacs package brainstorming prompt file.
Added curl communication backend with fallback support
- Added ollama-buddy-curl.el as separate backend implementation
- Implemented backend dispatcher system in ollama-buddy-core.el
- Updated all async functions to use backend dispatcher
- Added curl backend validation and testing functions
- Maintained full compatibility with existing network process backend
When building Emacs packages that communicate with external services, network connectivity can sometimes be a pain point. While Emacs’s built-in make-network-process works great in most cases, some users have encountered issues on certain systems or network configurations. That’s why now I have added a curl-based communication backend to give you an additional option, who knows, maybe it will solve your ollama communication issues!
The original ollama-buddy implementation relied exclusively on Emacs’s native network process functionality. While this works well for most users, I occasionally heard from users who experienced network process failures/flakiness on certain systems.
Rather than trying to work around these edge cases in the network process code, I took a different approach: I added curl as an alternative communication backend! This gives users a battle-tested, widely-available HTTP client as a fallback option.
Users can enable the curl backend with a simple customization:
(use-package ollama-buddy
;; :load-path "~/source/repos/ollama-buddy/ollama-buddy-mini"
:load-path "~/source/repos/ollama-buddy"
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:config
;; Load curl backend first
(require 'ollama-buddy-curl nil t)
;; Then set the backend
(setq ollama-buddy-communication-backend 'curl))and then switch backends from the chat buffer C-c e
The curl backend supports everything the network backend does:
- Real-time streaming responses
- Vision model support with image attachments
- File attachments and context
- All Ollama API parameters
- Multishot model sequences
If curl is selected but not available, the system automatically falls back to the network process with a helpful warning message.
From a user perspective, the backend choice is largely transparent. The main indicators are:
- Status line shows
[C]for curl or[N]for network - Process list shows
ollama-chat-curlvsollama-chat-streamprocesses - Curl backend shows “Curl Processing…” in status messages
Everything else - streaming behaviour, response formatting, error handling - works identically regardless of the backend.
Optimized Unicode escape function to fix blocking with large file attachments
- Fixed UI blocking when sending large attached files
- Used temp buffer with delete-char/insert instead of repeated concat calls
- Reduced processing time from minutes to milliseconds for large payloads
Full system prompt in the status bar replaced with a more meaningful simple role title
- Added system prompt metadata tracking with title, source, and timestamp registry
- Implemented automatic title extraction and unified completing-read interface
- Enhanced fabric/awesome prompt integration with proper metadata handling
- Improved transient menu organization and org-mode formatting with folding
- Added system prompt history display and better error handling for empty files
- Transient menu has been simplified and reorganised
Previously, the header status bar would show truncated system prompt text like [You are a helpful assistant wh...], making it difficult to quickly identify which prompt was active. Now, the display shows meaningful role titles with source indicators:
[F:Code Reviewer]- Fabric pattern[A:Linux Terminal]- Awesome ChatGPT prompt[U:Writing Assistant]- User-defined prompt
The system now intelligently extracts titles from prompt content by recognizing common patterns like “You are a…”, “Act as…”, or “I want you to act as…”. When these patterns aren’t found, it generates a concise title from the first few words.
Behind the scenes, Ollama Buddy now maintains a registry of all system prompts with their titles, sources, and timestamps. This enables new features like system prompt history viewing and better organization across Fabric patterns, Awesome ChatGPT prompts, and user-defined prompts.
The result is a cleaner interface that makes it immediately clear which role your AI assistant is currently embodying, without cluttering the status bar with long, truncated text.
Quite a bit of refactoring to generally make this project more maintainable and I have added a starter kit of user prompts.
- Color System Reworking
- Removed all model color-related functions and variables
- Removed dependency on
color.el - Replaced with
highlight-regexpand hashing to^font-lockfaces, so now using a more native built-in solutions for model colouring rather than shoe-horning in overlays.
- UI Improvements
- Simplified the display system by leveraging Org mode
- Added org-mode styling for output buffers
- Added
org-hide-emphasis-markersandorg-hide-leading-starssettings - Changed formatting to use Org markup instead of text properties
- Converted plain text headers to proper Org headings
- Replaced color properties with Org emphasis (bold)
- History Management Updates
- Streamlined history editing functionality
- Improved model-specific history editing
- Refactored history display and navigation
- System Prompts
- Added library of system prompts in these categories:
- analysis (3 prompts)
- coding (5 prompts)
- creative (3 prompts)
- documentation (3 prompts)
- emacs (10 prompts)
- general (3 prompts)
- technical (3 prompts)
- writing (3 prompts)
- Added library of system prompts in these categories:
Added user system prompts management
- You can now save, load and manage system prompts
- Created new transient menu for user system prompts (C-c s)
- Organized prompts by categories with org-mode format storage
- Supported prompt editing, listing, creation and deletion
- Updated key bindings to integrate with existing functionality
- Added prompts directory customization with defaults
This feature makes it easier to save, organize, and reuse your favorite system prompts when working with Ollama language models.
System prompts are special instructions that guide the behavior of language models. By setting effective system prompts, you can:
- Define the AI’s role (e.g., “You are a helpful programming assistant who explains code clearly”)
- Establish response formats
- Set the tone and style of responses
- Provide background knowledge for specific domains
The new ollama-buddy-user-prompts module organizes your system prompts in a clean, category-based system:
- Save your prompts - Store effective system prompts you’ve crafted for future use
- Categorize - Prompts are organized by domains like “coding,” “writing,” “technical,” etc.
- Quick access - Browse and load your prompt library with completion-based selection
- Edit in org-mode - All prompts are stored as org files with proper metadata
- Manage with ease - Create, edit, list, and delete prompts through a dedicated transient menu
The new functionality is accessible through the updated key binding C-c s, which opens a dedicated transient menu with these options:
- Save current (S) - Save your active system prompt
- Load prompt (L) - Choose a previously saved prompt
- Create new (N) - Start fresh with a new prompt
- List all Prompts (l) - View your entire prompt library
- Edit prompt (e) - Modify an existing prompt
- Delete prompt (d) - Remove prompts you no longer need
If you work frequently with Ollama models, you’ve likely discovered the power of well-crafted system prompts. They can dramatically improve the quality and consistency of responses. With this new management system, you can:
- Build a personal library of effective prompts
- Maintain context continuity across sessions
- Share prompts with teammates
- Refine your prompts over time
Added file attachment system for including documents in conversations
- Added file attachment support with configurable file size limits (10MB default) and supported file types
- Implemented session persistence for attachments in save/load functionality
- Added attachment context inclusion in prompts with proper token counting
- Created comprehensive attachment management commands:
- Attach files to conversations
- Show current attachments in dedicated buffer
- Detach specific files
- Clear all attachments
- Added Dired integration for bulk file attachment
- Included attachment menu in transient interface (C-c 1)
- Updated help text to document new attachment keybindings
- Enhanced context calculation to include attachment token usage
You can now seamlessly include text files, code, documentation, and more directly in your conversations with local AI models!
Simply use C-c C-a from the chat buffer to attach any file to your current conversation.
The attached files become part of your conversation context, allowing the AI to reference, analyze, or work with their contents directly.
The transient menu has also been updated with a new Attachment Menu
*File Attachments* a Attach file w Show attachments d Detach file 0 Clear all attachments
Your attachments aren’t just dumped into the conversation - they’re intelligently integrated:
- Token counting now includes attachment content, so you always know how much context you’re using
- Session persistence means your attachments are saved and restored when you save/load conversations
- File size limits (configurable, 10MB default) prevent accidentally overwhelming your context window
Managing attached files is intuitive with dedicated commands:
C-c C-w- View all current attachments in a nicely formatted org mode buffer, folded to each fileC-c C-d- Detach specific files when you no longer need themC-c 0- Clear all attachments at onceC-c 1- Access the full attachment menu via a transient interface
Working in Dired? No problem! You can attach files directly from your file browser:
- Mark multiple files and attach them all at once
- Attach the file at point with a single command
Use the configuration as follows:
(eval-after-load 'dired
'(progn
(define-key dired-mode-map (kbd "C-c C-a") #'ollama-buddy-dired-attach-marked-files)))Added context size management and monitoring
- Added configurable context sizes for popular models (llama3.2, mistral, qwen, etc.)
- Implemented real-time context usage display in status bar
- Can display in text or bar display types
- Added context size thresholds with visual warnings
- Added interactive commands for context management:
ollama-buddy-show-context-info: View all model context sizesollama-buddy-set-model-context-size: Manually configure model contextollama-buddy-toggle-context-percentage: Toggle context display
- Implemented context size validation before sending prompts
- Added token estimation and breakdown (history/system/current prompt)
- Added keybindings: C-c $ (set context), C-c % (toggle display), C-c C (show info)
- Updated status bar to show current/max context with fontification
I’ve added context window management and monitoring capabilities to Ollama Buddy!
This update helps you better understand and manage your model’s context usage, preventing errors and optimizing your conversations.
Enable it with the following:
(setq ollama-buddy-show-context-percentage t)After implementing these changes:
- Text mode: Shows
1024/4096style display - Bar mode (default): Shows
███████░░░░ 2048style display - Use
C-c 8to toggle between modes - The Text mode will change fontification based on your thresholds:
- Normal: regular fontification
- (85%+): underlined and bold
- (100%+): inverse video and bold
- The Bar mode will just fill up as normal
The progress bar will visually represent how much of the context window you’re using, making it easier to see at a glance when you’re approaching the limit.
Determining a model’s context size proved more complex than expected. While experimenting with parsing model info JSON, I discovered that context size information can be scattered across different fields. Rather than implementing a complex JSON parser (which may come later), I chose a pragmatic approach:
I created a new defcustom variable ollama-buddy-fallback-context-sizes that includes hard-coded values for popular Ollama models. The fallback mechanism is deliberately simple: substring matching followed by a sensible default of 4096 tokens.
(defcustom ollama-buddy-fallback-context-sizes
'(("llama3.2:1b" . 2048)
("llama3:8b" . 4096)
("tinyllama" . 2048)
("phi3:3.8b" . 4096)
("gemma3:1b" . 4096)
("gemma3:4b" . 8192)
("llama3.2:3b" . 8192)
("llama3.2:8b" . 8192)
("llama3.2:70b" . 8192)
("starcoder2:3b" . 8192)
("starcoder2:7b" . 8192)
("starcoder2:15b" . 8192)
("mistral:7b" . 8192)
("mistral:8x7b" . 32768)
("codellama:7b" . 8192)
("codellama:13b" . 8192)
("codellama:34b" . 8192)
("qwen2.5-coder:7b" . 8192)
("qwen2.5-coder:3b" . 8192)
("qwen3:0.6b" . 4096)
("qwen3:1.7b" . 8192)
("qwen3:4b" . 8192)
("qwen3:8b" . 8192)
("deepseek-r1:7b" . 8192)
("deepseek-r1:1.5b" . 4096))
"Mapping of model names to their default context sizes.
Used as a fallback when context size can't be determined from the API."
:type '(alist :key-type string :value-type integer)
:group 'ollama-buddy)This approach may not be perfectly accurate for all models, but it’s sufficient for getting the core functionality working. More importantly, as a defcustom, users can easily customize these values for complete accuracy with their specific models. Users can also set context values within the chat buffer through C-c C (Show Context Information) for each individual model if desired.
This design choice allowed me to focus on the essential features without getting stuck on complex context retrieval logic.
One final thing!, if the num_ctx: Context window size in tokens is set, then that number will also be taken into consideration. An assumption will be made that the model is honouring the context size requested and will incorporated into the context calculations accordingly.
For token counting, I’ve implemented a simple heuristic: each word (using string-split) is multiplied by 1.3. This follows commonly recommended approximations and works well enough in practice. While this isn’t currently configurable, I may add it as a customization option in the future.
The C-c C (Show Context Information) command is central to this feature. Rather than continuously monitoring context size while you type (which would be computationally expensive and potentially distracting), I’ve designed the system to calculate context on-demand when you choose.
Scenario 1: Paste-and-Send Approach
Let’s say you want to paste a large block of text into the chat buffer. You can simply:
- Paste your content
- Press the send keybinding
- If the context limit is exceeded, you’ll get a warning dialog asking whether to proceed anyway
Scenario 2: Preemptive Checking
For more control, you can check context usage before sending:
- Paste your content
- Run
C-c Cto see the current context breakdown - If the context looks too high, you have several options:
- Trim your current prompt
- Remove or simplify your system prompt
- Edit conversation history using Ollama Buddy’s history modification features
- Switch to a model with a larger context window
Scenario 3: Manage the Max History Length
Want tight control over context size without constantly monitoring the real-time display? Since conversation history is part of the context, you can simply limit ollama-buddy-max-history-length to control the total context size.
For example, when working with small context windows, set ollama-buddy-max-history-length to 1. This keeps only the last exchange (your prompt + model response), ensuring your context remains small and predictable, perfect for maintaining control without manual monitoring.
Scenario 4: Parameter num_ctx: Context window size in tokens
Simply set this parameter and off you go!
Given the potentially limiting nature of context management, I’ve set this feature to disabled by default.
But to enable set the following :
(setq ollama-buddy-show-context-percentage t)This means:
- Context checks won’t prevent sending prompts
- Context usage won’t appear in the status line
- However, calculations still run in the background, so
C-c C(Show Context Information) remains functional
As the feature matures and proves its value, I may enable it by default. For now, consider it an experimental addition that users can opt into.
The status bar now displays your current context usage in real-time. You’ll see a fraction showing used tokens versus the model’s maximum context size (e.g., “2048/8192”). The display automatically updates as your conversation grows.
Context usage changes fontification to help you stay within limits:
- Normal font: Normal usage (under 85%)
- Bold and Underlined: Approaching limit (85-100%)
- Inversed: At or exceeding limit (100%+)
Before sending prompts that exceed the context limit, Ollama Buddy now warns you and asks for confirmation. This prevents unexpected errors and helps you manage long conversations more effectively.
There are now three new interactive commands:
C-c $ - Set Model Context Size. Manually configure context sizes for custom or fine-tuned models.
C-c % - Toggle Context Display. Show or hide the context percentage in the status bar.
C-c C - Show Context Information. View a detailed breakdown of:
- All model context sizes
- Current token usage by category (history, system prompt, current prompt)
- Percentage usage
The system estimates token counts for:
- Conversation history: All previous messages
- System prompts: Your custom instructions
- Current input: The message you’re about to send
This gives you a complete picture of your context usage before hitting send.
The context monitoring is not enabled by default.
- Sorted model names alphabetically in intro message
- Removed multishot writing to register name letters
For some reason, when I moved the .ollama folder to an external disk, the models returned with api/tags were inconsistent, which meant it broke consistent letter assignment. I’m not sure why this happened, but it is probably sensible to sort the models alphabetically anyway, as this has the benefit of naturally grouping together model families.
I also removed the multishot feature of writing to the associated model letter. Now that I have to accommodate more than 26 models, incorporating them into the single-letter Emacs register system is all but impossible. I suspect this feature was not much used, and if you think about it, it wouldn’t have worked anyway with multiple model shots, as the register letter associated with the model would just show the most recent response. Due to these factors, I think I should remove this feature. If someone wants it back, I will probably have to design a bespoke version fully incorporated into the ollama-buddy system, as I can’t think of any other Emacs mechanism that could accommodate this.
Fix model reference error exceeding 26 models #15
Update ollama-buddy to handle more than 26 models by using prefixed combinations for model references beyond ‘z’. This prevents errors in create-intro-message when the local server hosts a large number of models.
Added the following to recommended models:
- qwen3:0.6b
- qwen3:1.7b
- qwen3:4b
- qwen3:8b
and fixed pull model
Refactored model prefixing again so that when using only ollama models no prefix is applied and is only applied when online LLMs are selected (for example claude, chatGPT e.t.c)
I think this makes more sense and is cleaner for I suspect the majority who may use this package are probably more interested in just using ollama models and the prefix will probably be a bit confusing.
This could be a bit of a breaking change once again I’m afraid for those ollama users that have switched and are now familiar with prefixing “o:”, sorry!
Added vision support for those ollama models that can support it!
Image files are now detected within a prompt and then processed if a model can support vision processing. Here’s a quick overview of how it works:
- Configuration: Users can configure the application to enable vision support and specify which models and image formats are supported. Vision support is enabled by default.
- Image Detection: When a prompt is submitted, the system automatically detects any image files referenced in the prompt.
- Vision Processing: If the model supports vision, the detected images are processed in relation to the defined prompt. Note that the detection of a model being vision capable is defined in
ollama-buddy-vision-modelsand can be adjusted as required. - In addition, a menu item has been added to the custom ollama buddy menu :
[I] Analyze an Image
When selected, it will allow you to describe a chosen image. At some stage, I may allow integration into dired, which would be pretty neat. :)
Added model unloading functionality to free system resources
- Add unload capability for individual models via the model management UI
- Create keyboard shortcut (C-c C-u) for quick unloading of all models
- Display running model count and unload buttons in model management buffer
Large language models consume significant RAM and GPU memory while loaded. Until now, there wasn’t an easy way to reclaim these resources without restarting the Ollama server entirely. This new functionality allows you to:
- Free up GPU memory when you’re done with your LLM sessions
- Switch between resource-intensive tasks more fluidly
- Manage multiple models more efficiently on machines with limited resources
- Avoid having to restart the Ollama server just to clear memory
There are several ways to unload models with the new functionality:
- Unload All Models: Press
C-c C-uto unload all running models at once (with confirmation) - Model Management Interface: Access the model management interface with
C-c Wwhere you’ll find:- A counter showing how many models are currently running
- An “Unload All” button to free all models at once
- Individual “Unload” buttons next to each running model
- Quick Access in Management Buffer: When in the model management buffer, simply press
uto unload all models
The unloading happens asynchronously in the background, with clear status indicators so you can see when the operation completes.
- Display modified parameters in token stats
Enhanced the token statistics section to include any modified parameters, providing a clearer insight into the active configurations. This update helps in debugging and understanding the runtime environment.
Added Reasoning/Thinking section visibility toggle functionality
- Introduced the ability to hide reasoning/thinking sections during AI responses, making the chat output cleaner and more focused on final results
- Added a new customizable variable
ollama-buddy-hide-reasoning(default: nil) which controls visibility of reasoning sections - Added
ollama-buddy-reasoning-markersto configure marker pairs that encapsulate reasoning sections (supports multiple formats like <think></think> or ----) - Added
ollama-buddy-toggle-reasoning-visibilityinteractive command to switch visibility on/off - Added keybinding
C-c Vfor toggling reasoning visibility in chat buffer - Added transient menu option “V” for toggling reasoning visibility
- When reasoning is hidden, a status message shows which section is being processed (e.g., “Think…” or custom marker names)
- Reasoning sections are automatically detected during streaming responses
- Header line now indicates when reasoning is hidden with “REASONING HIDDEN” text
- All changes preserve streaming response functionality while providing cleaner output
This feature is particularly useful when working with AI models that output their “chain of thought” or reasoning process before providing the final answer, allowing users to focus on the end results while still having the option to see the full reasoning when needed.
Added Grok support
Integration is very similar to other remote AIs:
(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:custom
(ollama-buddy-grok-api-key
(auth-source-pick-first-password :host "ollama-buddy-grok" :user "apikey"))
:config
(require 'ollama-buddy-grok nil t))Fixed utf-8 encoding stream response issues from remote LLMs.
Finished the remote LLM decoupling process, meaning that the core ollama-buddy logic is now not dependent on any remote LLM, and each remote LLM package is self-contained and functions as a unique extension.
Refactored model prefixing logic and cleaned up
- Standardized model prefixing by introducing distinct prefixes for Ollama (
o:), OpenAI (a:), Claude (c:), and Gemini (g:) models. - Centralized functions to get full model names with prefixes across different model types.
- Removed redundant and unused variables related to model management.
Note that there may be some breaking changes here especially regarding session recall as all models will now have a prefix to uniquely identify their type. For ollama recall, just edit the session files to prepend the ollama prefix of “o:”
Added Gemini integration!
As with the Claude and ChatGPT integration, you will need to add something similar to them in your configuration. I currently have the following set up to enable access to the remote LLMs:
(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:custom
(ollama-buddy-openai-api-key
(auth-source-pick-first-password :host "ollama-buddy-openai" :user "apikey"))
(ollama-buddy-claude-api-key
(auth-source-pick-first-password :host "ollama-buddy-claude" :user "apikey"))
(ollama-buddy-gemini-api-key
(auth-source-pick-first-password :host "ollama-buddy-gemini" :user "apikey"))
:config
(require 'ollama-buddy-openai nil t)
(require 'ollama-buddy-claude nil t)
(require 'ollama-buddy-gemini nil t))Also with the previous update all the latest model names will be pulled, so there should be a full comprehensive list for each of the main remote AI LLMs!
See README-features.org for a deeper delve into each feature as it was added.
- Minimal Setup
If desired, the following will get you going! (of course, have ollama running with some models loaded).
(use-package ollama-buddy
:ensure t
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper))- Interactive Command Menu
- Quick-access menu with single-key commands (
M-x ollama-buddy-menu) - Quickly define your own menu using
defcustomwith dynamic adaptable menu - Menu Presets easily definable in text files
- Quickly switch between LLM models with no configuration required
- Send text from any Emacs buffer
- Each model is uniquely colored to enhance visual feedback
- Switch between basic and advanced interface levels (
C-c A)
- Quick-access menu with single-key commands (
- Role and Session Management
- Create, switch, save and manage role-specific command menu configurations
- Create, switch, save and manage sessions to provide chat/model based context
- Prompt history/context with the ability to clear, turn on and off and save as part of a session file
- Edit conversation history with intuitive keybindings
- Smart Model Management
- Models can be assigned to individual menu commands
- Intelligent model fallback
- Real-time model availability monitoring
- Easy model switching during sessions
- Parameter Control
- Comprehensive parameter management for all Ollama API options
- Temperature control for adjusting AI creativity
- Command-specific parameter customization for optimized interactions
- Visual parameter interface showing active and modified values
- System Prompt Support
- Set persistent system prompts for guiding AI responses
- Command-specific system prompts for tailored interactions
- Visual indicators for active system prompts
- AI Operations
- Code refactoring with context awareness
- Automatic git commit message generation
- Code explanation and documentation
- Text operations (proofreading, conciseness, dictionary lookups)
- Custom prompt support for flexibility
- Conversation Tools
- Conversation history tracking and navigation
- Real-time token usage tracking and statistics
- Multi-model comparison (send prompts to multiple models)
- Markdown to Org conversion for better formatting
- Export conversations in various formats
- Lightweight
- Single package file
- A minified version
ollama-buddy-mini(200 lines) is available - No external dependencies (curl not used)
Ollama Buddy now includes a transient-based menu system to improve usability and streamline interactions. Yes, I originally stated that I would never do it, but I think it compliments my crafted simple textual menu and the fact that I have now defaulted the main chat interface to a simple menu.
This can give the user more options for configuration, they can use the chat in advanced mode where the keybindings are presented in situ, or a more minimal basic setup where the transient menu can be activated. For my use-package definition I current have the following set up, with the two styles of menus sitting alongside each other :
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu)The new menu provides an organized interface for accessing the assistant’s core functions, including chat, model management, roles, and Fabric patterns. This post provides an overview of the features available in the Ollama Buddy transient menus.
Yes that’s right also fabric patterns!, I have decided to add in auto syncing of the patterns directory in https://github.com/danielmiessler/fabric
Simply I pull the patterns directory which contain prompt guidance for a range of different topics and then push them through a completing read to set the ollama-buddy system prompt, so a special set of curated prompts can now be applied right in the ollama-buddy chat!
Anyways, here is a description of the transient menu system.
The transient menu in Ollama Buddy leverages Emacs’ transient.el package (the same technology behind Magit’s popular interface) to create a hierarchical, discoverable menu system. This approach transforms the user experience from memorizing numerous keybindings to navigating through logical groups of commands with clear descriptions.
The main transient menu can be accessed with the keybinding C-c O when in an Ollama Buddy chat buffer. You can also call it via M-x ollama-buddy-transient-menu from anywhere in Emacs.
When called, the main transient menu appears at the bottom of your Emacs frame, organized into logical sections with descriptive prefixes. Here’s what you’ll see:
|o(Y)o| Ollama Buddy [Chat] [Prompts] [Model] [Roles & Patterns] o Open Chat l Send Region W Manage Models R Switch Roles O Commands s Set System Prompt m Switch Model E Create New Role RET Send Prompt C-s Show System v View Model Status D Open Roles Directory h Help/Menu r Reset System i Show Model Info f Fabric Patterns k Kill/Cancel b Ollama Buddy Menu M Multishot t OpenAI Integration x Toggle Streaming [Display Options] [History] [Sessions] [Parameters] A Toggle Interface Level H Toggle History N New Session P Edit Parameter B Toggle Debug Mode X Clear History L Load Session G Display Parameters T Toggle Token Display J Edit History S Save Session I Parameter Help u Token Stats Q List Sessions K Reset Parameters U Token Usage Graph Z Delete Session F Toggle Params in Header C-o Toggle Markdown->Org p Parameter Profiles c Toggle Model Colors V Toggle reasoning visibility
This visual layout makes it easy to discover and access the full range of Ollama Buddy’s functionality. Let’s explore each section in detail.
This section contains the core interaction commands:
- Open Chat (o): Opens the Ollama Buddy chat buffer
- Commands (O): Opens a submenu with specialized commands
- Send Prompt (RET): Sends the current prompt to the model
- Help/Menu (h): Displays the help assistant with usage tips
- Kill/Cancel Request (k): Cancels the current ongoing request
- Toggle Streaming (x): Toggle streaming on and off only for ollama models
These commands help you manage and send prompts:
- Send Region (l): Sends the selected region as a prompt
- Set System Prompt (s): Sets the current prompt as a system prompt
- Show System Prompt (C-s): Displays the current system prompt
- Reset System Prompt (r): Resets the system prompt to default
- Ollama Buddy Menu (b): Opens the classic custom menu interface
Commands for model management:
- Manage Models: Info, Pull, Copy, Delete for each model, plus Import GGUF Model and Pull any model from ollama library
- Switch Model (m): Changes the active LLM
- View Model Status (v): Shows status of all available models
- Show Model Info (i): Displays detailed information about the current model
- Multishot (M): Sends the same prompt to multiple models
These commands help manage roles and use fabric patterns:
- Switch Roles (R): Switch to a different predefined role
- Create New Role (E): Create a new role interactively
- Open Roles Directory (D): Open the directory containing role definitions
- Fabric Patterns (f): Opens the submenu for Fabric patterns
- OpenAI Integration (t): Opens menu specifically for OpenAI/ChatGPT
When you select the Fabric Patterns option, you’ll see a submenu like this:
Fabric Patterns (42 available, last synced: 2025-03-18 14:30) [Actions] [Sync] [Categories] [Navigation] s Send with Pattern S Sync Latest u Universal Patterns q Back to Main Menu p Set as System P Populate Cache c Code Patterns l List All Patterns I Initial Setup w Writing Patterns v View Pattern Details a Analysis Patterns
Commands to customize the display:
- Toggle Interface Level (A): Switch between basic and advanced interfaces
- Toggle Debug Mode (B): Enable/disable JSON debug information
- Toggle Token Display (T): Show/hide token usage statistics
- Display Token Stats (U): Show detailed token usage information
- Toggle Markdown->Org (C-o): Enable/disable conversion to Org format
- Toggle Model Colors (c): Enable/disable model-specific colors
- Toggle reasoning visibility (V): Hide/show reasoning/thinking sections
- Token Usage Graph (g): Display a visual graph of token usage
Commands for managing conversation history:
- Toggle History (H): Enable/disable conversation history
- Clear History (X): Clear the current history
- Edit History (J): Edit the history in a buffer
Commands for session management:
- New Session (N): Start a new session
- Load Session (L): Load a saved session
- Save Session (S): Save the current session
- List Sessions (Q): List all available sessions
- Delete Session (Z): Delete a saved session
Commands for managing model parameters:
- Edit Parameter (P): Opens a submenu to edit specific parameters
- Display Parameters (G): Show current parameter settings
- Parameter Help (I): Display help information about parameters
- Reset Parameters (K): Reset parameters to defaults
- Toggle Params in Header (F): Show/hide parameters in header
- Parameter Profiles (p): Opens the parameter profiles submenu
When you select the Edit Parameter option, you’ll see a comprehensive submenu of all available parameters:
Parameters
[Generation] [More Generation] [Mirostat]
t Temperature f Frequency Penalty M Mirostat Mode
k Top K s Presence Penalty T Mirostat Tau
p Top P n Repeat Last N E Mirostat Eta
m Min P x Stop Sequences
y Typical P l Penalize Newline
r Repeat Penalty
[Resource] [More Resource] [Memory]
c Num Ctx P Num Predict m Use MMAP
b Num Batch S Seed L Use MLOCK
g Num GPU N NUMA C Num Thread
G Main GPU V Low VRAM
K Num Keep o Vocab Only
[Profiles] [Actions]
d Default Profile D Display All
a Creative Profile R Reset All
e Precise Profile H Help
A All Profiles F Toggle Display in Header
q Back to Main Menu
Ollama Buddy includes predefined parameter profiles that can be applied with a single command. When you select “Parameter Profiles” from the main menu, you’ll see:
Parameter Profiles Current modified parameters: temperature, top_k, top_p [Available Profiles] d Default c Creative p Precise [Actions] q Back to Main Menu
The Commands submenu provides quick access to specialized operations:
Ollama Buddy Commands [Code Operations] [Language Operations] [Pattern-based] [Custom] r Refactor Code l Dictionary Lookup f Fabric Patterns C Custom Prompt d Describe Code s Synonym Lookup u Universal Patterns m Minibuffer Prompt g Git Commit Message p Proofread Text c Code Patterns [Actions] q Back to Main Menu
For experienced users who prefer direct keybindings, all transient menu functions can also be accessed through keybindings with the prefix of your choice (or C-c O when in the chat minibuffer) followed by the key shown in the menu. For example:
C-c O s- Set system promptC-c O m- Switch modelC-c O P- Open parameter menu
The transient menu can be customized by modifying the transient-define-prefix definitions in the package. You can add, remove, or rearrange commands to suit your workflow.
Note that all the demos are in real time.
Also these videos will all be uploaded to a youtube channel:
https://www.youtube.com/@OllamaBuddyforEmacs
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
(setq font-general “Source Code Pro 15”)
theme : doom-oceanic-next
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
- Starting with model : llama3.2:1b
- Show menu activation C-c o
ollama-buddy-menu - [o] Open chat buffer
- PROMPT:: why is the sky blue?
- C-c C-c to send from chat buffer
- From this demo README file, select the following text and send to chat buffer: What is the capital of France?
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
- C-c m to swap to differing models from chat buffer
- Select models from the intro message
- Swap models from the transient menu
- PROMPT:: how many fingers do I have?
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
the quick brown fox jumps over the lazy dog
- Select individual words for dictionary menu items
- Select whole sentence and convert to uppercase
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
004 Coding - Ollama Buddy - Writing a Hello World! program with increasingly advanced models #emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
- PROMPT:: can you write a hello world program in Ada?
- switch models to the following and check differing output:
- tinyllama:latest
- qwen2.5-coder:3b
- qwen2.5-coder:7b
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
In fairy-tales, witches always wear silly black hats and black coats, and they ride on broomsticks. But this is not a fairy-tale. This is about REAL WITCHES.
Lets change roles into a Buffy preset and have some fun!
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
- Cordelia burn
- Giles… yawn!
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
- PROMPT:: how many fingers do I have?
- send to multiple models, any difference in output?
- Also they are now available in the equivalent letter named registers, so lets pull those registers!
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
Show the changing of roles and how they affect the menu and hence the commands available.
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
Enter a few queries to test the system, then navigate back through your previous inputs, switch models, and resubmit to compare results.
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
- Ask:
What is the capital of France? and of Italy?
- Turn off history
and of Germany?
- Turn on history
and of Germany?
- Save Session
- Restart Emacs
- Load Session
and of Sweden?
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
Set the system message to: You must always respond in a single sentence. Now ask the following: Tell me why Emacs is so great! Tell me about black holes clear the system message and ask again, the responses should now be more verbose!!
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
PROMPT: What is the capital of France? Multishot to abcdefghij
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
Display Token Stats Display Token Usage Graph
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
Select a passage from 20,000 and push
as a poet as a gaslighter as a drunk person
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
switch to use qwen2.5-coder:3b
Select User System Prompts
List, and move down through the buffer
Set as system prompt/load
Select Elisp Debugging Guide
Show the system prompt in a buffer
Run checking overy my/rsync in Emacs-DIYer
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
switch to use qwen2.5-coder:3b
Attach the ollama-buddy Makefile
can you tell me about this file?
Attach another file, curl-tests.sh
Show attachments
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
Demonstrating the Emacs package ollama-buddy, providing a convenient way to integrate Ollama’s local LLM capabilities.
switch to use gemma3:4b
Open the image with the text
Copy the link, paste into the chat buffer and then lets see if the text is correctly extracted!
https://melpa.org/#/ollama-buddy https://github.com/captainflasmr/ollama-buddy
#emacs #ollama
- Ollama installed and running locally
- Emacs 26.1 or later
(use-package ollama-buddy
:ensure t
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper))Clone this repository:
git clone https://github.com/captainflasmr/ollama-buddy.gitWith the option to add your own user keybinding for the ollama-buddy-menu
(add-to-list 'load-path "path/to/ollama-buddy")
(require 'ollama-buddy)
(global-set-key (kbd "C-c o") #'ollama-buddy-menu)
(global-set-key (kbd "C-c O") #'ollama-buddy-transient-menu-wrapper)OR
(use-package ollama-buddy
:load-path "path/to/ollama-buddy"
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper))- Usage
- Start your Ollama server locally with
ollama serve - In Emacs,
M-x ollama-buddy-menuor a user defined keybindingC-c oto open up the ollama buddy menu - Select [o] to open and jump to the chat buffer
- Read the AI assistants greeting and off you go!
| Key | Action |
|---|---|
| C-c C-c | Send prompt |
| C-c k | Cancel request |
| C-c m | Change model |
| C-c h | Show help |
| C-c i | Show model info |
| C-c U | Show token statistics |
| C-c U | Show token graph |
| C-c V | Toggle reasoning visibility |
| C-c C-s | Show system prompt |
| C-u C-c C-c | Set system prompt |
| C-u C-u C-c C-c | Clear system prompt |
| M-p / M-n | Browse prompt history |
| C-c n / C-c p | Navigate between prompts |
| C-c N | New session |
| C-c L | Load session |
| C-c S | Save session |
| C-c Y | List sessions |
| C-c W | Delete session |
| C-c H | Toggle history |
| C-c X | Clear history |
| C-c E | Edit history |
| C-c l | Prompt to multiple models |
| C-c P | Edit parameters |
| C-c G | Show parameters |
| C-c I | Show parameter help |
| C-c K | Reset parameters |
| C-c D | Toggle JSON debug mode |
| C-c T | Toggle token display |
| C-c Z | Toggle parameters |
| C-c C-o | Toggle Markdown/Org format |
| C-c A | Toggle interface level (basic/advanced) |
The default menu offers the following menu items:
| Key | Action | Description |
|---|---|---|
| o | Open chat buffer | Open chat buffer |
| m | Swap model | Swap model |
| v | View model status | View model status |
| l | Send region | Send region |
| R | Switch roles | Switch roles |
| N | Create new role | Create new role |
| D | Open roles directory | Open roles directory |
| r | Refactor code | Refactor code |
| g | Git commit message | Git commit message |
| c | Describe code | Describe code |
| d | Dictionary Lookup | Dictionary Lookup |
| n | Word synonym | Word synonym |
| p | Proofread text | Proofread text |
| e | Custom prompt | Custom prompt |
| i | Minibuffer Prompt | Minibuffer Prompt |
| K | Delete session | Delete session |
| q | Quit | Quit |
Lets quickly get started by adding that little command you frequently use on text, lets add it to the ollama buddy menu!
First, lets clarify some concepts.
In ollama-buddy, a “command” is essentially a predefined action that appears as an option in the menu. Each command has:
- A key (the button you press to invoke it)
- A description (what shows up in the menu)
- An action (what happens when you select it)
- Optional settings like prompts, model selection, etc.
A “role” is a collection of commands. Think of it as a preset configuration or profile that defines which commands are available in your menu and how they behave.
There are two types of prompts that might be causing confusion:
- Prompt (
:promptproperty): This is the instruction or question that gets prepended to your selected text. For example, if your prompt is “Fill in the missing Turkish letters in:” and your selected text is “Merhaba nasilsin”, the LLM will receive “Fill in the missing Turkish letters in: Merhaba nasilsin”. - System Prompt (
:systemproperty): This is a behind-the-scenes instruction that guides the LLM’s overall behavior but isn’t directly part of your query. It’s like telling the AI “Here’s how I want you to approach this task” before you give it the actual task.
Let’s set up a simple, consistent command for filling in missing Turkish letters, there are two main ways.
The first will add a menu item to the existing menu, and the second will define a new role, which will be a collection of menu items. In the example, I have added only a single menu item, but you can add as many as you like. If you want, you can set up an entire menu system in the Role file that could be just Turkey specific.
(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:config
(setq ollama-buddy-default-model "o:tinyllama:latest")
(add-to-list 'ollama-buddy-command-definitions
'(turkish-letters
:key ?t ; Press 't' in the menu to select this
:description "Fix Turkish Letters" ; What shows in the menu
:model "o:deepseek-r1:7b" ; Choose an appropriate model
:prompt "Fill in the missing Turkish letters in the following text:"
:system "You are an expert in Turkish language. Your task is to correctly add any missing Turkish-specific letters (ç, ğ, ı, ö, ş, ü) to text that may be missing them. Only correct the spelling by adding proper Turkish letters - do not change any words or add any commentary. Return only the corrected text."
:action (lambda () (ollama-buddy--send-with-command 'turkish-letters)))))- Add the code block above to your Emacs configuration (tweaking the use-package command as desired for your local configuration and setting the :model you wish to use)
- Restart Emacs or evaluate the code
- Select some Turkish text in any buffer
- Press
C-c oto open the ollama-buddy menu - Press
tto run your “Fix Turkish Letters” command
- Create the file
~/.emacs.d/ollama-buddy-presets/ollama-buddy--preset__turkish.elwith the code block below:
(require 'ollama-buddy)
(setq ollama-buddy-command-definitions
'(
;; Standard commands (always include these)
(open-chat
:key ?o
:description "Open chat buffer"
:action ollama-buddy--open-chat)
(show-models
:key ?v
:description "View model status"
:action ollama-buddy-show-model-status)
(switch-role
:key ?R
:description "Switch roles"
:action ollama-buddy-roles-switch-role)
(create-role
:key ?E
:description "Create new role"
:action ollama-buddy-role-creator-create-new-role)
(open-roles-directory
:key ?D
:description "Open roles directory"
:action ollama-buddy-roles-open-directory)
(swap-model
:key ?m
:description "Swap model"
:action ollama-buddy--swap-model)
(help
:key ?h
:description "Help"
:action ollama-buddy--menu-help-assistant)
;; Your custom Turkish command
(turkish-letters
:key ?t
:description "Fix Turkish Letters"
:model "o:deepseek-r1:7b"
:prompt "Fill in the missing Turkish letters in the following text:"
:system "You are an expert in Turkish language. Your task is to correctly add any missing Turkish-specific letters (ç, ğ, ı, ö, ş, ü) to text that may be missing them. Only correct the spelling by adding proper Turkish letters - do not change any words or add any commentary. Return only the corrected text."
:action (lambda () (ollama-buddy--send-with-command 'turkish-letters)))
(send-region
:key ?l
:description "Send region"
:action (lambda ()
(let* ((selected-text (when (use-region-p)
(buffer-substring-no-properties
(region-beginning) (region-end)))))
(when (not selected-text)
(user-error "This command requires selected text"))
(ollama-buddy--open-chat)
(insert selected-text))))
))- In Emacs, press
C-c oto open the ollama-buddy menu - Press
Rto switch roles - Select “turkish” from the list
- Now your menu will be simplified and focused on Turkish helpers
- Select some Turkish text, open the menu again (
C-c o), and presst
Let’s break down what’s happening in the command definition:
(turkish-letters
:key ?t ; The key to press in the menu
:description "Fix Turkish Letters" ; What appears in the menu
:model "o:deepseek-r1:7b"
:prompt "Fill in the missing Turkish letters in the following text:" ; Instruction sent to the LLM
:system "You are an expert in Turkish language..." ; Background instruction for the LLM
:action (lambda () (ollama-buddy--send-with-command 'turkish-letters))) ; Function to execute:promptis what gets added before your selected text. It tells the AI what to do with your text.:systemgives the AI context about its role and specific behavioral guidelines. This helps ensure it gives exactly the type of response you want.
The prompt is for the specific task, while the system message shapes the AI’s overall approach. Together, they ensure consistent, specialized behavior.
Roles in Ollama Buddy allow you to create different configurations of commands and models for specific use cases. This tutorial will guide you through setting up roles, creating custom menus, and effectively using them in your workflow.
A role is essentially a preset configuration that defines:
- Which commands are available in your menu
- What models to use for specific commands
- What prompts and system messages to use
- Any special parameters for optimization
For example, you might have a “programmer” role focused on coding tasks and a “writer” role with writing-focused commands.
The file naming convention is critical to understand how roles, preset files, and menu configurations work together:
- Required filename format:
ollama-buddy--preset__ROLE-NAME.el- The double underscore
__separates the prefix from your role name - The role name portion becomes the identifier shown when switching roles
- Example:
ollama-buddy--preset__programmer.elcreates a role named “programmer”
- The double underscore
This naming convention is how Ollama Buddy discovers and identifies role files in your roles directory. When you run ollama-buddy-roles-switch-role, the system:
- Scans the
ollama-buddy-roles-directoryfor files matching the pattern - Extracts the role name from each filename (the part after
__) - Presents these names in the role selection interface
- When selected, loads the corresponding file which redefines
ollama-buddy-command-definitions - This redefinition immediately changes the available commands in your Ollama Buddy menu
The relationship chain works like this:
ollama-buddy--preset__ROLE-NAME.el → Defines ollama-buddy-command-definitions → Controls menu content
When creating roles using the interactive role creator (C-c E), this naming convention is automatically handled for you. When creating roles manually, you must follow this pattern for Ollama Buddy to recognize your role files correctly.
You can locate your roles directory with:
;; Check where your roles are stored
(message ollama-buddy-roles-directory)
;; Or open the directory directly
M-x ollama-buddy-roles-open-directoryBy default, this is set to ~/.emacs.d/ollama-buddy-presets/, but you can customize it:
(setq ollama-buddy-roles-directory "/your/custom/path/to/presets")There are two ways to create custom roles:
The most user-friendly approach:
- Press
C-c Eor runM-x ollama-buddy-role-creator-create-new-role - Enter a name for your role (e.g., “programmer”)
- For each command you want to add:
- Specify a command name (e.g., “refactor-code”)
- Choose a key shortcut for the menu
- Add a description
- Optionally specify a model
- Optionally add prompt prefixes and system messages
For more advanced customization, create role files manually:
- Create a file named
ollama-buddy--preset__your-role-name.elin yourollama-buddy-roles-directory - Structure your file like this:
;; ollama-buddy preset for role: programmer
(require 'ollama-buddy)
(setq ollama-buddy-command-definitions
'(
;; Standard commands - always include these
(open-chat
:key ?o
:description "Open chat buffer"
:action ollama-buddy--open-chat)
(show-models
:key ?v
:description "View model status"
:action ollama-buddy-show-model-status)
(switch-role
:key ?R
:description "Switch roles"
:action ollama-buddy-roles-switch-role)
(create-role
:key ?E
:description "Create new role"
:action ollama-buddy-role-creator-create-new-role)
(open-roles-directory
:key ?D
:description "Open roles directory"
:action ollama-buddy-roles-open-directory)
;; Custom commands for this role
(refactor-code
:key ?r
:description "Refactor code"
:model "codellama:7b"
:prompt "Refactor this code to improve readability and efficiency:"
:system "You are an expert software engineer who improves code quality while maintaining functionality."
:action (lambda () (ollama-buddy--send-with-command 'refactor-code)))
(explain-code
:key ?e
:description "Explain code"
:model "deepseek-r1:7b"
:prompt "Explain what this code does in detail:"
:system "You are a programming teacher who explains code clearly and thoroughly."
:action (lambda () (ollama-buddy--send-with-command 'explain-code)))
(git-commit
:key ?g
:description "Git commit message"
:prompt "Write a concise git commit message for these changes:"
:system "You are a version control expert who creates clear, concise commit messages."
:action (lambda () (ollama-buddy--send-with-command 'git-commit)))
))To switch between roles:
- Press
C-c Ror runM-x ollama-buddy-roles-switch-role - Select a role from the completion list
- The menu will update with the commands defined in that role
You can also switch roles from within the Ollama Buddy menu by pressing ‘R’.
Assign specific models to commands for optimal performance:
(ollama-buddy-add-model-to-menu-entry 'refactor-code "codellama:7b")Optimize parameters for specific commands:
(ollama-buddy-add-parameters-to-command 'refactor-code
'temperature 0.2
'top_p 0.7
'repeat_penalty 1.3)Add entirely new commands to your menu:
(ollama-buddy-update-menu-entry 'my-new-command
:key ?z
:description "My new awesome command"
:prompt "Here is what I want you to do:"
:system "You are an expert system specialized in this task."
:action (lambda () (ollama-buddy--send-with-command 'my-new-command)))Here’s a complete example of setting up a writing-focused role:
;; ollama-buddy preset for role: writer
(require 'ollama-buddy)
(setq ollama-buddy-command-definitions
'(
;; Standard commands
(open-chat
:key ?o
:description "Open chat buffer"
:action ollama-buddy--open-chat)
(show-models
:key ?v
:description "View model status"
:action ollama-buddy-show-model-status)
(switch-role
:key ?R
:description "Switch roles"
:action ollama-buddy-roles-switch-role)
(create-role
:key ?E
:description "Create new role"
:action ollama-buddy-role-creator-create-new-role)
(open-roles-directory
:key ?D
:description "Open roles directory"
:action ollama-buddy-roles-open-directory)
;; Writing-focused commands
(summarize
:key ?s
:description "Summarize text"
:prompt "Summarize the following text in a concise manner:"
:system "You are an expert at extracting the key points from any text."
:action (lambda () (ollama-buddy--send-with-command 'summarize)))
(proofread
:key ?p
:description "Proofread text"
:model "deepseek-r1:7b"
:prompt "Proofread the following text and correct any errors:"
:system "You are a professional editor who identifies and corrects grammar, spelling, and style errors."
:action (lambda () (ollama-buddy--send-with-command 'proofread)))
(rewrite
:key ?r
:description "Rewrite text"
:prompt "Rewrite the following text to improve clarity and flow:"
:system "You are a skilled writer who can improve any text while preserving its meaning."
:action (lambda () (ollama-buddy--send-with-command 'rewrite)))
(brainstorm
:key ?b
:description "Brainstorm ideas"
:model "llama3.2:3b"
:prompt "Generate creative ideas related to the following topic:"
:parameters ((temperature . 1.0) (top_p . 0.95))
:action (lambda () (ollama-buddy--send-with-command 'brainstorm)))
))Save this as ollama-buddy--preset__writer.el in your ollama-buddy-roles-directory.
- Group related commands: Create roles around specific workflows or tasks
- Match models to tasks: Use lightweight models for simple tasks and more powerful models for complex ones
- Customize system prompts: Craft specific system prompts to guide the model for each command
- Use the roles directory: Press
C-c Dto quickly access and manage your role files - Create specialized roles: Consider roles for programming, writing, translation, or domain-specific knowledge
The ollama-buddy chat buffer is powered by org-mode, providing enhanced readability and structure. Conversations automatically format user prompts and AI responses with org-mode headings, making them easier to navigate and utilize.
Benefits include:
- Outlining and heading navigation
- Org export capabilities
- Source code fontification
Toggle Markdown to Org conversion with C-c C-o.
You can choose between two interface levels depending on your preference:
- Basic Interface: Shows minimal commands for new users
- Advanced Interface: Shows all available commands and features
Set your preferred interface level:
(setq ollama-buddy-interface-level 'basic) ; or 'advancedBy default, the menu is set to Basic. Switch between levels during your session with C-c A.
Ollama Buddy maintains context between your interactions by:
- Tracking conversation history between prompts and responses
- Sending previous messages to Ollama for improved contextual responses
- Displaying a history counter in the status line showing conversation length
- Providing configurable history length limits to control memory usage
Control this feature with:
;; Enable/disable conversation history (default: t)
(setq ollama-buddy-history-enabled t)
;; Set maximum conversation pairs to remember (default: 10)
(setq ollama-buddy-max-history-length 10)
;; Show/hide the history counter in the header line (default: t)
(setq ollama-buddy-show-history-indicator t)History-related commands:
C-c H: Toggle history tracking on/offC-c X: Clear the current conversation historyC-c E: Edit conversation history (universal argument to edit specific model)
Comprehensive parameter management gives you complete control over your Ollama model’s behavior through API options:
- All Parameters - Set any custom option for the Ollama LLM at runtime
- Smart Parameter Management: Only modified parameters are sent to Ollama, preserving defaults
- Visual Parameter Interface: Clear display showing which parameters are active with highlighting
Access parameter management through keyboard shortcuts:
C-c P- Edit a parameterC-c G- Display current parametersC-c I- Show parameter helpC-c K- Reset parameters to defaults
You can define specific parameter sets for each command in the menu, enabling optimization for particular use cases:
;; Update properties and parameters at once
(ollama-buddy-update-command-with-params 'describe-code
:model "codellama:latest"
:parameters '((temperature . 0.3) (top_p . 0.8)))This feature is particularly useful for:
- Code-related tasks: Lower temperature for more deterministic code generation
- Creative writing: Higher temperature for more varied and creative outputs
- Technical explanations: Balanced settings for clear, accurate explanations
- Summarization tasks: Custom parameters to control verbosity and focus
Ollama Buddy supports system prompts, allowing you to set and manage system-level instructions for your AI interactions:
- Set a system prompt: Use
C-u C-c C-cto designate any user prompt as a system prompt - Clear system prompt: Use
C-u C-u C-c C-cto clear the system prompt - View system prompt: Use
C-c C-sto display the current system prompt
System prompts remain active across user queries, providing better control over conversation context. The status bar displays an “S” indicator when a system prompt is active.
You can also define system prompts per command:
(ollama-buddy-update-menu-entry
'refactor-code
:model "qwen2.5-coder:7b"
:system "You are an expert software engineer who improves code and only mainly using the principles exhibited by Ada")Real-time token tracking helps you monitor usage and performance:
- Track token counts, rates, and usage history
- Display token usage statistics (
C-c tor menu option) - Toggle token stats display after responses
- Real-time updates via timer
With the multishot mode, you can send a prompt to multiple models in sequence and compare their responses:
- Models are assigned letters for quick selection (e.g.,
(a) mistral,(b) gemini) - Use
C-c Mto initiate a multishot sequence - Status updates track progress during multishot execution
To use multishot mode:
C-c Mto start a multishot session- Type a sequence of model letters (e.g.,
abcto use models a, b, and c) - The selected models process the prompt one by one
Roles in Ollama Buddy are essentially profiles tailored to specific tasks:
- Store custom roles in
ollama-buddy-roles-directory(default:~/.emacs.d/ollama-buddy-presets/) - Switch between roles with
M-x ollama-buddy-roles-switch-roleor menu optionR - Create custom roles with
M-x ollama-buddy-role-creator-create-new-roleor menu optionN - Open role directory with
M-x ollama-buddy-roles-open-directoryor menu optionD
Preconfigured presets available:
- ollama-buddy–preset__buffy.el
- ollama-buddy–preset__default.el
- ollama-buddy–preset__emacs.el
- ollama-buddy–preset__developer.el
- ollama-buddy–preset__janeway.el
- ollama-buddy–preset__translator.el
- ollama-buddy–preset__writer.el
Session management allows you to save conversations and restore them with relevant context:
- Save session with
ollama-buddy-sessions-saveorC-c S - Load session with
ollama-buddy-sessions-loadorC-c L - List sessions with
ollama-buddy-sessions-listorC-c Y - Delete session with
ollama-buddy-sessions-deleteorC-c K - New session with
ollama-buddy-sessions-neworC-c E
Sessions preserve model-specific chat history to prevent context contamination across different models.
Prompts are integrated into the Emacs history mechanism and persist across sessions:
- Use
M-pto navigate prompt history in the chat buffer - Use
M-p/M-nwithin the minibuffer to insert previous prompts
Ollama Buddy provides a flexible menu system that can be easily customized to match your workflow. The menu is built from ollama-buddy-command-definitions, which you can modify or extend in your Emacs configuration.
Each menu item is defined using a property list with these key attributes:
(command-name
:key ?k ; Character for menu selection
:description "desc" ; Menu item description
:model "model-name" ; Specific Ollama model (optional)
:prompt "prompt" ; System prompt (optional)
:system "system" ; System prompt (optional)
:parameters ((param . value)) ; Command-specific parameters (optional)
:action function) ; Command implementationYou can add new commands to ollama-buddy-command-definitions in your config:
;; Add a single new command
(add-to-list 'ollama-buddy-command-definitions
'(pirate
:key ?i
:description "R Matey!"
:model "mistral:latest"
:prompt "Translate the following as if I was a pirate:"
:action (lambda () (ollama-buddy--send-with-command 'pirate))))
;; Incorporate into a use-package
(use-package ollama-buddy
:load-path "path/to/ollama-buddy"
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
(add-to-list 'ollama-buddy-command-definitions
'(pirate
:key ?i
:description "R Matey!"
:model "mistral:latest"
:prompt "Translate the following as if I was a pirate:"
:action (lambda () (ollama-buddy--send-with-command 'pirate))))
:custom ollama-buddy-default-model "llama:latest")
;; Add multiple commands at once
(setq ollama-buddy-command-definitions
(append ollama-buddy-command-definitions
'((summarize
:key ?u
:description "Summarize text"
:model "tinyllama:latest"
:prompt "Provide a brief summary:"
:action (lambda ()
(ollama-buddy--send-with-command 'summarize)))
(translate-spanish
:key ?t
:description "Translate to Spanish"
:model "mistral:latest"
:prompt "Translate this text to Spanish:"
:action (lambda ()
(ollama-buddy--send-with-command 'translate-spanish))))))You can create a minimal configuration by defining only the commands you need:
;; Minimal setup with just essential commands
(setq ollama-buddy-command-definitions
'((send-basic
:key ?l
:description "Send Basic Region"
:action (lambda () (ollama-buddy--send-with-command 'send-basic)))
(quick-define
:key ?d
:description "Define word"
:model "tinyllama:latest"
:prompt "Define this word:"
:action (lambda ()
(ollama-buddy--send-with-command 'quick-define)))
(quit
:key ?q
:description "Quit"
:model nil
:action (lambda ()
(message "Quit Ollama Shell menu.")))))- Choose unique keys for menu items
- Match models to task complexity (small models for quick tasks)
- Use clear, descriptive names
| Property | Description | Required |
|---|---|---|
| :key | Single character for menu selection | Yes |
| :description | Menu item description | Yes |
| :model | Specific Ollama model to use | No |
| :prompt | Static system prompt | No |
| :system | System-level instruction | No |
| :parameters | Command-specific parameters | No |
| :action | Function implementing the command | Yes |
- Model Selection and Fallback Logic
You can associate specific commands defined in the menu with an Ollama LLM to optimize performance for different tasks. For example, if speed is a priority over accuracy, such as when retrieving synonyms, you might use a lightweight model like TinyLlama or a 1B–3B model. On the other hand, for tasks that require higher precision, like code refactoring, a more capable model such as Qwen-Coder 7B can be assigned to the “refactor” command on the buddy menu system.
Since this package enables seamless model switching through Ollama, the buddy menu can present a list of commands, each linked to an appropriate model. All Ollama interactions share the same chat buffer, ensuring that menu selections remain consistent. Additionally, the status bar on the header line and the prompt itself indicate the currently active model.
Ollama Buddy also includes a model selection mechanism with a fallback system to ensure commands execute smoothly, even if the preferred model is unavailable.
Commands in ollama-buddy-command-definitions can specify preferred models using the :model property. This allows optimizing different commands for specific models:
(defcustom ollama-buddy-command-definitions
'((refactor-code
:key ?r
:description "Refactor code"
:model "qwen-coder:latest"
:prompt "refactor the following code:")
(git-commit
:key ?g
:description "Git commit message"
:model "tinyllama:latest"
:prompt "write a concise git commit message for the following:")
(send-region
:key ?l
:description "Send region"
:model "llama:latest"))
...)When :model is nil, the command will use whatever model is currently set as ollama-buddy-default-model.
When executing a command, the model selection follows this fallback chain:
- Command-specific model (
:modelproperty) - Current model (
ollama-buddy-default-model) - User selection from available models
(setq ollama-buddy-default-model "llama:latest")When a fallback occurs, Ollama Buddy provides clear feedback:
- The header line shows which model is being used
- If using a fallback model, an orange warning appears showing both the requested and actual model
- The model status can be viewed using the “View model status” command (v key)
(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:custom
;; Set default model
(ollama-buddy-default-model "llama3:latest")
;; Set interface level (basic or advanced)
(ollama-buddy-interface-level 'advanced)
;; Enable model colors
(ollama-buddy-enable-model-colors t)
;; Set menu columns
(ollama-buddy-menu-columns 3))(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:custom
(ollama-buddy-host "http://<remote-server>")
(ollama-buddy-port 11400))(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:config
(ollama-buddy-add-model-to-menu-entry 'dictionary-lookup "tinyllama:latest")
(ollama-buddy-add-model-to-menu-entry 'synonym "tinyllama:latest"))(use-package ollama-buddy
:bind
("C-c o" . ollama-buddy-menu)
("C-c O" . ollama-buddy-transient-menu-wrapper)
:config
(ollama-buddy-update-menu-entry
'refactor-code
:model "qwen2.5-coder:7b"
:system "You are an expert software engineer who improves code using Ada principles"))| Custom variable | Description |
|---|---|
| ollama-buddy-awesome-categorize-prompts | Whether to categorize prompts based on common keywords. |
| ollama-buddy-awesome-prompts-file | Filename containing the prompts within the repository. |
| ollama-buddy-grok-api-endpoint | Endpoint for Grok chat completions API. |
| ollama-buddy-remote-models | List of available remote models. |
| ollama-buddy-command-definitions | Comprehensive command definitions for Ollama Buddy. |
| ollama-buddy-enable-model-colors | Whether to show model colors. |
| ollama-buddy-interface-level | Level of interface complexity to display. |
| ollama-buddy-show-params-in-header | Whether to show modified parameters in the header line. |
| ollama-buddy-debug-mode | When non-nil, show raw JSON messages in a debug buffer. |
| ollama-buddy-mode | Non-nil if Ollama-Buddy mode is enabled. |
| ollama-buddy-grok-marker-prefix | Prefix used to identify Grok models in the ollama-buddy interface. |
| ollama-buddy-reasoning-markers | List of marker pairs that encapsulate reasoning/thinking sections. |
| ollama-buddy-awesome-update-on-startup | Whether to automatically update prompts when Emacs starts. |
| ollama-buddy-context-bar-width | Width of the context progress bar in characters. |
| ollama-buddy-sessions-directory | Directory containing ollama-buddy session files. |
| ollama-buddy-history-model-view-mode-map | Keymap for model-specific history viewing mode. |
| ollama-buddy-menu-columns | Number of columns to display in the Ollama Buddy menu. |
| ollama-buddy-gemini-api-endpoint | Endpoint format for Google Gemini API. |
| ollama-buddy-fabric-pattern-categories | List of pattern categories to focus on when listing patterns. |
| ollama-buddy-status-update-interval | Interval in seconds to update the status line with background operations. |
| ollama-buddy-fabric-update-on-startup | Whether to automatically update patterns when Emacs starts. |
| ollama-buddy-available-models | List of available models to pull from Ollama Hub. |
| ollama-buddy-grok-max-tokens | Maximum number of tokens to generate in the response. |
| ollama-buddy-claude-max-tokens | Maximum number of tokens to generate in the response. |
| ollama-buddy-openai-marker-prefix | Prefix to indicate that a model is from OpenAI rather than Ollama. |
| ollama-buddy-history-enabled | Whether to use conversation history in Ollama requests. |
| ollama-buddy-grok-api-key | API key for accessing Grok services. |
| ollama-buddy-grok-default-model | Default Grok model to use. |
| ollama-buddy-claude-marker-prefix | Prefix used to identify Claude models in the model list. |
| ollama-buddy-roles-directory | Directory containing ollama-buddy role preset files. |
| ollama-buddy-history-view-mode-map | Keymap for history viewing mode. |
| ollama-buddy-fabric-local-dir | Local directory where Fabric patterns will be stored. |
| ollama-buddy-show-history-indicator | Whether to show the history indicator in the header line. |
| ollama-buddy-context-size-thresholds | Thresholds for context usage warnings. |
| ollama-buddy-mode-map | Keymap for ollama-buddy mode. |
| ollama-buddy-marker-prefix | Prefix used to identify Ollama models in the ollama-buddy interface. |
| ollama-buddy-fabric-patterns-subdir | Subdirectory within the Fabric repo containing the patterns. |
| ollama-buddy-context-display-type | How to display context usage in the status bar. |
| ollama-buddy-awesome-repo-url | URL of the Awesome ChatGPT Prompts GitHub repository. |
| ollama-buddy-openai-api-endpoint | Endpoint for OpenAI chat completions API. |
| ollama-buddy-vision-models | List of models known to support vision capabilities. |
| ollama-buddy-params-active | Currently active values for Ollama API parameters. |
| ollama-buddy-openai-default-model | Default OpenAI model to use. |
| ollama-buddy-claude-default-model | Default Claude model to use. |
| ollama-buddy-params-modified | Set of parameters that have been explicitly modified by the user. |
| ollama-buddy-gemini-temperature | Temperature setting for Gemini requests (0.0-1.0). |
| ollama-buddy-current-session-name | The name of the currently loaded session. |
| ollama-buddy-host | Host where Ollama server is running. |
| ollama-buddy-image-formats | List of regular expressions matching supported image file formats. |
| ollama-buddy-streaming-enabled | Whether to use streaming mode for responses. |
| ollama-buddy-max-history-length | Maximum number of message pairs to keep in conversation history. |
| ollama-buddy-gemini-default-model | Default Gemini model to use. |
| ollama-buddy-default-model | Default Ollama model to use. |
| ollama-buddy-claude-temperature | Temperature setting for Claude requests (0.0-1.0). |
| ollama-buddy-vision-enabled | Whether to enable vision support for models that support it. |
| ollama-buddy-awesome-prompt-variable | |
| ollama-buddy-display-token-stats | Whether to display token usage statistics in responses. |
| ollama-buddy-show-context-percentage | Whether to show context percentage in the status bar. |
| ollama-buddy-claude-api-endpoint | Endpoint for Anthropic Claude API. |
| ollama-buddy-gemini-api-key | API key for accessing Google Gemini services. |
| ollama-buddy-openai-max-tokens | Maximum number of tokens to generate in the response. |
| ollama-buddy-claude-api-key | API key for accessing Anthropic Claude services. |
| ollama-buddy-params-profiles | Predefined parameter profiles for different usage scenarios. |
| ollama-buddy-awesome-local-dir | Local directory where Awesome ChatGPT Prompts will be stored. |
| ollama-buddy-params-defaults | Default values for Ollama API parameters. |
| ollama-buddy-mode-line-segment | Mode line segment for Ollama Buddy. |
| ollama-buddy-grok-temperature | Temperature setting for Grok requests (0.0-1.0). |
| ollama-buddy-port | Port where Ollama server is running. |
| ollama-buddy-default-register | Default register to store the current response when not in multishot mode. |
| ollama-buddy-fabric-repo-url | URL of the Fabric GitHub repository. |
| ollama-buddy-gemini-max-tokens | Maximum number of tokens to generate in the response. |
| ollama-buddy-modelfile-directory | Directory for storing temporary Modelfiles. |
| ollama-buddy-fallback-context-sizes | Mapping of model names to their default context sizes. |
| ollama-buddy-context-bar-chars | Characters used to draw the context progress bar. |
| ollama-buddy-openai-temperature | Temperature setting for OpenAI requests (0.0-2.0). |
| ollama-buddy-convert-markdown-to-org | Whether to automatically convert markdown to `org-mode’ format in responses. |
| ollama-buddy-mode-hook | Hook run after entering or leaving `ollama-buddy-mode’. |
| ollama-buddy-gemini-marker-prefix | Prefix used to identify Gemini models in the ollama-buddy interface. |
| ollama-buddy-hide-reasoning | When non-nil, hide reasoning/thinking blocks from the stream output. |
| ollama-buddy-openai-api-key | API key for accessing OpenAI services. |
| Command | Description |
|---|---|
| ollama-buddy-open-info | Open the Info manual for the ollama-buddy package. |
| ollama-buddy-dired-import-gguf | Import the GGUF file at point in Dired into Ollama. |
| ollama-buddy-toggle-markdown-conversion | Toggle automatic conversion of markdown to ‘org-mode’ format. |
| ollama-buddy-fabric-send | Apply a Fabric pattern to the selected text and send to Ollama. |
| ollama-buddy-reset-system-prompt | Reset the system prompt to default (none). |
| ollama-buddy-beginning-of-prompt | Move point to the beginning of the current prompt. |
| ollama-buddy-toggle-streaming | Toggle streaming mode for Ollama responses. |
| ollama-buddy-toggle-reasoning-visibility | Toggle visibility of reasoning/thinking sections in responses. |
| ollama-buddy-history-toggle-edit-mode | Toggle between viewing and editing modes for history. |
| ollama-buddy-mode | Minor mode for ollama-buddy keybindings. |
| ollama-buddy-roles-switch-role | Switch to a different ollama-buddy role. |
| ollama-buddy-pull-model | Pull or update MODEL from Ollama Hub asynchronously. |
| ollama-buddy-history-edit | View and edit the conversation history in a buffer. |
| ollama-buddy-previous-history | Navigate to previous item in prompt history. |
| ollama-buddy-toggle-params-in-header | Toggle display of modified parameters in the header line. |
| ollama-buddy-display-token-graph | Display a visual graph of token usage statistics. |
| ollama-buddy-toggle-token-display | Toggle display of token statistics after each response. |
| ollama-buddy-show-raw-model-info | Retrieve and display raw JSON information about the current default MODEL. |
| ollama-buddy-history-cancel | Cancel editing the history. |
| ollama-buddy-sessions-load | Load an Ollama Buddy session with improved org file handling. |
| ollama-buddy-transient-menu-wrapper | Wrapper function for safely loading the Ollama Buddy transient menu. |
| ollama-buddy-sessions-list | Display a list of saved sessions. |
| ollama-buddy-toggle-context-percentage | Toggle display of context percentage in the status bar. |
| ollama-buddy-transient-awesome-menu | Awesome ChatGPT Prompts for ollama-buddy. |
| ollama-buddy-show-context-info | Show detailed information about context sizes for all models. |
| ollama-buddy-fabric-set-system-prompt | Set the system prompt to a Fabric pattern without sending a request. |
| ollama-buddy-params-help | Display help for Ollama parameters. |
| ollama-buddy-awesome-send | Apply an Awesome ChatGPT Prompt to the selected text and send to Ollama. |
| ollama-buddy-sessions-new | Start a new session by clearing history and buffer. |
| ollama-buddy-show-system-prompt | Display the current system prompt in a buffer. |
| ollama-buddy-awesome-show-prompts-menu | Show a transient menu of Awesome ChatGPT Prompt organized by category. |
| ollama-buddy-awesome-populate-prompts | Populate the list of available prompt from the local repository. |
| ollama-buddy-awesome-set-system-prompt | Set the system prompt to an Awesome ChatGPT Prompt without sending a request. |
| ollama-buddy-params-edit | Edit a specific parameter PARAM interactively. |
| ollama-buddy-awesome-list-prompts | Display a list of available Awesome ChatGPT Prompt. |
| ollama-buddy-transient-parameter-menu | Parameter menu for Ollama Buddy. |
| ollama-buddy-toggle-debug-mode | Toggle display of raw JSON messages in a debug buffer. |
| ollama-buddy-awesome-show-prompt | Display the full content of a prompt with FORMATTED-NAME. |
| ollama-buddy-toggle-model-colors | Toggle the use of model-specific colors in ollama-buddy. |
| ollama-buddy-awesome-setup | Set up the ollama-buddy-awesome package. |
| ollama-buddy-fabric-show-pattern | Display the full content of a PATTERN. |
| ollama-buddy-fabric-sync-patterns | Sync the latest patterns from the Fabric GitHub repository. |
| ollama-buddy-reset-all-prompts | Reset both system prompt and suffix to default (none). |
| ollama-buddy-fabric-list-patterns | Display a list of available Fabric patterns with descriptions. |
| ollama-buddy-history-search | Search through the prompt history using a ‘completing-read’ interface. |
| ollama-buddy-toggle-history | Toggle conversation history on/off. |
| ollama-buddy-display-token-stats | Display token usage statistics. |
| ollama-buddy-role-creator-create-new-role | Create a new role interactively. |
| ollama-buddy-roles-open-directory | Open the ollama-buddy roles directory in Dired. |
| ollama-buddy-params-reset | Reset all parameters to default values and clear modification tracking. |
| ollama-buddy-menu | Display Ollama Buddy menu with support for prefixed model references. |
| ollama-buddy-unload-all-models | Unload all currently running Ollama models to free up resources. |
| ollama-buddy-history-toggle-edit-model | Toggle between viewing and editing modes for MODEL history. |
| ollama-buddy-transient-menu | Ollama Buddy main menu. |
| ollama-buddy-fabric-setup | Set up the ollama-buddy-fabric package. |
| ollama-buddy-manage-models | Update the model management interface to include unload capabilities. |
| ollama-buddy-history-edit-model | Edit the conversation history for a specific MODEL. |
| ollama-buddy-roles-create-directory | Create the ollama-buddy roles directory if it doesn’t exist. |
| ollama-buddy-history-save-model | Save the edited history for MODEL back to variable. |
| ollama-buddy-toggle-context-display-type | Toggle between text and bar display for context usage. |
| ollama-buddy-set-max-history-length | Set the maximum number of message pairs to keep in conversation history. |
| ollama-buddy-reset-suffix | Reset the suffix to default (none). |
| ollama-buddy-show-model-status | Display status of models referenced in command definitions with color coding. |
| ollama-buddy-next-history | Navigate to next item in prompt history. |
| ollama-buddy-clear-history | Clear the conversation history. |
| ollama-buddy-awesome-sync-prompts | Sync the latest prompt from the Awesome ChatGPT Prompt GitHub repository. |
| ollama-buddy-set-model-context-size | Manually set the context size for MODEL to SIZE. |
| ollama-buddy-sessions-delete | Delete an Ollama Buddy session. |
| ollama-buddy-transient-profile-menu | Parameter profiles menu for Ollama Buddy. |
| ollama-buddy-set-suffix | Set the current prompt as a suffix. |
| ollama-buddy-set-system-prompt | Set the current prompt as a system prompt. |
| ollama-buddy-history-save | Save the edited history back to ‘ollama-buddy–conversation-history-by-model’. |
| ollama-buddy-import-gguf-file | Import a GGUF file at FILE-PATH into Ollama. |
| ollama-buddy-transient-commands-menu | Commands menu for Ollama Buddy. |
| ollama-buddy-toggle-interface-level | Toggle between basic and advanced interface levels. |
| ollama-buddy-transient-fabric-menu | Fabric patterns menu for Ollama Buddy. |
| ollama-buddy-params-display | Display the current Ollama parameter settings. |
| ollama-buddy-fabric-populate-patterns | Populate the list of available patterns from the local repository. |
| ollama-buddy-sessions-save | Save the current Ollama Buddy session. |
All interactive functions available in ollama-buddy are listed in the original README.
| API Endpoint | Method | Purpose | Supported? | will ollama-buddy support? |
|---|---|---|---|---|
/api/chat | POST | Send chat messages to a model | Yes | |
/api/generate | POST | Generate text without chat context | No | Probably not, as chat covers |
/api/delete | DELETE | Deletes available models | Yes | |
/api/tags | GET | List available models | Yes | |
/api/pull | POST | Pull a model from Ollama library | Yes | |
/api/push | POST | Push a model to the Ollama library | No | No |
/api/create | POST | Create a model from a Modelfile | Yes | |
/api/show | POST | Show model information | Yes | |
/api/ps | GET | List running models | Yes | |
/api/copy | POST | Copy a model | Yes | |
/api/embed | POST | Generate embeddings from text | No | Probably when I can figure it out |
/api/version | GET | Get Ollama version | Yes | |
| Cancel requests | Custom | Terminate ongoing request | Yes |
| Parameter | Supported | Notes | will ollama-buddy support? |
|---|---|---|---|
model | Yes | Can be specified per command or globally | |
prompt | Yes | Core feature with extensive prompt handling | |
suffix | Yes/No | Can be set with ollama-buddy-set-suffix | |
images | No | Unlikely, this is Emacs! | |
format | No | Maybe | |
options | Yes | Full implementation of all model parameters | |
system | Yes | Can be set with ollama-buddy-set-system-prompt | |
template | No | Maybe | |
stream | Yes | Toggleable with ollama-buddy-toggle-streaming | |
raw | No | Might as well | |
keep_alive | No | Uses default Ollama behavior | Yes |
| Parameter Group | Parameters | Supported |
|---|---|---|
| Temperature Controls | temperature, top_k, top_p, min_p, typical_p | Yes |
| Repetition Controls | repeat_last_n, repeat_penalty, presence_penalty, frequency_penalty | Yes |
| Advanced Sampling | mirostat, mirostat_tau, mirostat_eta, penalize_newline, stop | Yes |
| Resource Management | num_keep, seed, num_predict, num_ctx, num_batch | Yes |
| Hardware Optimization | numa, num_gpu, main_gpu, low_vram, vocab_only, use_mmap, use_mlock, num_thread | Yes |
| TODO | DOING | DONE |
|---|---|---|
| GGUF to be asynchronous | Kill online AI process with cancel request | Claude request and connection can sometimes take time and it blocks! |
| Better attachment signalling | allow switch to remote LLMs even when ollama not running | |
| Modifying a parameter internal defaults would flag as not modified | Vision image with spaces not processed as image | |
| No streaming with hiding reasoning corrupts output | Save system prompts as part of session | |
| Pushing hundreds or even thousands of lines to the chat buffer takes a while | chat ollama names to include prefix | |
| Check tags caching | Still some encoding issues on returns from online AI | |
| keybindings in ollama chat buffer are reserved | Conversation history to display with Claude/ChatGPT | |
| I think some hard coded Claude models currently don’t work | ||
| register contents to be converted to org, currently it is in markdown | ||
C-a beginning-of-line to jump to prompt start like comint | ||
| As shell generally C-n to end of list to clear prompt history | ||
| Add Texinfo to MELPA recipe | ||
| System prompt not cleared on new session | ||
| Remove fabric from Commands transient menu | ||
| External AI to copy to register | ||
| First request with no chat buffer fails |
| TODO | DOING | DONE |
|---|---|---|
| Autoload intelligently? | Add option to hide reasoning in chat buffer | Unload all models |
| Core API - Embeddings | Core Params - suffix | check context size of current model when querying |
| Core Params - format | Show token parameters as part of token stats | |
| Core Params - template | Autogenerate session save name | |
| Core Params - raw | Add gemini? | |
| Test on Windows | Pull ChatGPT and Claude model names | |
| Merge history buffer functionality | ||
| Show incoming ollama messages | ||
| Add chatGPT awesome prompt support | ||
| Add fabric pattern/prompt support | ||
| ChatGPT support | ||
| Claude support | ||
| Add new ChatGPT 4.1 | ||
| Method for securely storing API keys | ||
| Core API - Delete Model | ||
| Core API - Copy a model | ||
| Core API - stream | ||
| Core API - Pull/Info/Pull/Delete a model | ||
| Core API - Show running models | ||
| Core API - Create a model from a GGUF file | ||
| Add transient menu | ||
| History editing | ||
| Core API - options - parameter editing | ||
| Core API - show model information | ||
| Core API - system prompt | ||
| Chat buffer to use org-mode | ||
| Chat export options | ||
| Core Params - options:temperature | ||
| Managing Sessions | ||
| Sparse minified version | ||
| Role-Based Menu Preset System | ||
| Add to MELPA | ||
| Enhancing Menu Clarity | ||
| Real time token tracking | ||
| Managing chat history and context |
I think this is only for generate, but I have quietly added it in.
To the best of my knowledge, there are currently a few Emacs packages related to Ollama, though the ecosystem is still relatively young:
- llm.el (by Andrew Hyatt)
- A more general LLM interface package that supports Ollama as one of its backends
- GitHub: https://github.com/ahyatt/llm
- Provides a more abstracted approach to interacting with language models
- Supports multiple backends including Ollama, OpenAI, and others
- gptel (by Karthik Chikmagalur)
- While primarily designed for ChatGPT and other online services, it has experimental Ollama support
- GitHub: https://github.com/karthink/gptel
- Offers a more integrated chat buffer experience
- Has some basic Ollama integration, though it’s not the primary focus
- chatgpt-shell (by xenodium)
- Primarily designed for ChatGPT, but has some exploration of local model support
- GitHub: https://github.com/xenodium/chatgpt-shell
- Not specifically Ollama-focused, but interesting for comparison
- ellama (by s-kostyaev)
- A comprehensive Emacs package for interacting with local LLMs through Ollama
- GitHub: https://github.com/s-kostyaev/ellama
- Features deep org-mode integration and extensive prompt templates
- Offers streaming responses and structured interaction patterns
- More complex but feature-rich approach to local LLM integration
Let’s compare ollama-buddy to the existing solutions:
- llm.el
- Pros:
- Provides a generic LLM interface
- Supports multiple backends
- More abstracted and potentially more extensible
ollama-buddyis more:- Directly focused on Ollama
- Lightweight and Ollama-native
- Provides a more interactive, menu-driven approach
- Simpler to set up for Ollama specifically
- Pros:
- gptel
- Pros:
- Sophisticated chat buffer interface
- Active development
- Good overall UX
ollama-buddydifferentiates by:- Being purpose-built for Ollama
- Offering a more flexible, function-oriented approach
- Providing a quick, lightweight interaction model
- Having a minimal, focused design
- Pros:
- chatgpt-shell
- Pros:
- Mature shell-based interaction model
- Rich interaction capabilities
ollama-buddystands out by:- Being specifically designed for Ollama
- Offering a simpler, more direct interaction model
- Providing a quick menu-based interface
- Having minimal dependencies
- Pros:
- ellama
- Pros:
- Tight integration with Emacs org-mode
- Extensive built-in prompt templates
- Support for streaming responses
- Good documentation and examples
ollama-buddydiffers by:- Having a simpler, more streamlined setup process
- Providing a more lightweight, menu-driven interface
- Focusing on quick, direct interactions from any buffer
- Having minimal dependencies and configuration requirements
- Pros:
Report issues on the GitHub Issues page
If the local ollama server has more than 26 models the create-intro-message function produce an error due to the fact that alphabet letter are just 26.
in the ollama cli when using models with vision support (ex. ollama run gemma3:4b) you can say something like “extract the text from this image: /home/user/Downloads/screenshot.png”) then the ollama cli will detect this valid path and pass it to the gemma model as a file parameter somehow and the model will then be able to analize the actual file, when using ollama-buddy this is not possible
I am simply trying to add a new role which should by default add the missing turkish letters to any of the paragraph of text I send to it. However, ollama-buddy’s Role stuff is quite confusing to me. How do I set it up? I see no explanation about what I am doing when I try to add a role like that.
I just updated ollama-buddy’s source to its latest version, here’s my use-package config:
(use-package ollama-buddy
:if (file-directory-p "/home/user/.local/gits/emacs/ollama-buddy")
:vc (:url "file:///home/user/.local/gits/emacs/ollama-buddy"
:rev "870233e5d83da133b42fa7e9f57fcd6909f74502")
:config
(setq ollama-buddy-current-model "gemma3:12b-it-qat")
(setq ollama-buddy-menu-columns 4)
(setq ollama-buddy-host "localhost")
(setq ollama-buddy-port 11434)
(setq ollama-buddy-connection-check-interval 1))After byte-compiling, I cannot use ollama-buddy. When I send a prompt to my ollama backend, I constantly get Debugger entered–Lisp error: (excessive-lisp-nesting 1601). Older versions of ollama-buddy have been working without such an error.
As for context size, all I want the model to report it’s max size it can handle, what was the context size received, what it considered as input context, what it dropped and output context size for each call. Visually, all I was want is red, green amber colouring to know where it cut off(red), was ok(green) or slightly exceeded the context size(amber). whatever the model reports on the metrics(if at all, it does track that on a per API call)
Thanks for the package, I’m slowly getting into the grove of using it regularly.
I have a laptop without graphics cards for LLM work. I get about 2-3 tokens/sec. The newer models seem to be doing reasoning/thinking text output that are huge and takes up time. Is there a way to toggle it off during the session or through config on a per model basis. Either as a binary switch or a config parameter.
While at it, how do I set the context length to larger sizes than the default one and how do I check if it’s respected? On a per model basis, when you report the token stats, would you consider plugging in info about temp, top-k, context length that was active during the session?
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch
- Commit your changes
- Open a pull request
- Ollama for making local LLM inference accessible
- Emacs community for continuous inspiration
