Own your AI coding memory. Continue anywhere.
Capture, organize, and restore your AI coding conversations across every tool.
Every day, developers have rich conversations with AI coding assistants — discussing architecture, debugging issues, making design decisions. But these conversations are ephemeral:
- Close the terminal? Context gone.
- Switch from Claude Code to Cursor? Start over.
- Come back after a week? Forgot everything.
- Compact kicked in? Details lost forever.
Your AI conversations are valuable decision-making records, not disposable chat logs.
Memora runs quietly in the background, capturing conversations from all your AI coding tools and organizing them by project. When you need to continue work — in any tool — Memora gives you back your full context.
| Tool | Capture Method | Status |
|---|---|---|
| Claude Code | Hook (real-time) | Supported |
| Codex CLI | Hook (real-time) | Supported |
| OpenCode | Hook + SQLite | Supported |
| Cursor | SQLite polling | Supported |
The core differentiator. Select conversations, git history, and project files from any combination of tools and sessions, then generate a new workspace with complete context:
- Git clone of your codebase
MEMORA_CONTEXT.mdwith full conversation history (raw, uncompressed)- Automatic
CLAUDE.md/AGENTS.mdreferences so AI tools read the context - Works with any AI coding tool
View conversations and git commits side by side in a synchronized timeline. See what was discussed and what was built, together.
AI-powered project analysis: recent outcomes, active discussions, and unlanded ideas — generated from your conversation and git history.
Organizes project knowledge into layers:
| Layer | Name | Example |
|---|---|---|
| L1 | Persona | User preferences, coding style |
| L2 | Context | CLAUDE.md, project rules |
| L3 | Strategy | Design docs, plans |
| L4 | Execution | Tasks, TODOs |
| L5 | Dialogue | Conversations (the process) |
| L6 | Outcome | Git commits (the results) |
Go back to any point in your project's history and create a new branch — with full conversation context from that moment, plus optional "hindsight" about what happened after.
- macOS (Apple Silicon)
- One or more supported AI coding tools
Download the latest DMG from Releases.
- Open the DMG
- Drag Memora to Applications
- Launch Memora — it will ask to install hooks into your AI tools
- Start coding with any supported tool — Memora captures automatically
On first launch, Memora will:
- Ask permission to install hooks into detected AI tools
- Import existing conversation history
- Show an interactive guide of the interface
Hook CLI → ~/.memora/events.jsonl + ~/.memora/raw/
↓
SQLite (index)
↓
Flutter Desktop UI
- Frontend: Flutter (macOS desktop)
- Backend: Rust (3 crates: memora-core, memora-ingest, memora-hook)
- Bridge: flutter_rust_bridge v2.12
- Data: All data stays local. Nothing is uploaded anywhere.
Memora is 100% local. Your conversations, code, and data never leave your machine. There is no cloud sync, no telemetry, no analytics.
Memora supports English and Chinese (中文). Switch in Settings.
Leon Xu (云梦泽) — EastWorld ltd.
Memora is proprietary software. All rights reserved.
每天，开发者都与 AI 编程助手进行丰富的对话——讨论架构、调试问题、做设计决策。但这些对话是短暂的：
- 关闭终端？上下文消失。
- 从 Claude Code 切换到 Cursor？从头开始。
- 一周后回来？什么都忘了。
- 压缩（Compact）触发了？细节永远丢失。
你的 AI 对话是有价值的决策记录，而不是一次性的聊天日志。
Memora 在后台安静运行，从你所有的 AI 编程工具中捕获对话，并按项目整理。当你需要继续工作时——无论使用哪个工具——Memora 都能还原你的完整上下文。
| 工具 | 捕获方式 | 状态 |
|---|---|---|
| Claude Code | Hook（实时） | 已支持 |
| Codex CLI | Hook（实时） | 已支持 |
| OpenCode | Hook + SQLite | 已支持 |
| Cursor | SQLite 轮询 | 已支持 |
核心差异化功能。从任意工具和 session 的组合中选择对话、Git 历史和项目文件，然后生成一个具有完整上下文的新工作区：
- 代码库的 Git clone
MEMORA_CONTEXT.md，包含完整对话历史（原始、未压缩）- 自动添加
CLAUDE.md/AGENTS.md引用，使 AI 工具读取上下文 - 适用于任何 AI 编程工具
在同步的时间线上并排查看对话和 Git commit。同时看到讨论了什么和构建了什么。
AI 驱动的项目分析：近期成果、活跃讨论和未落地的想法——基于你的对话和 Git 历史生成。
将项目知识组织为不同层级：
| 层级 | 名称 | 示例 |
|---|---|---|
| L1 | 人格 | 用户偏好、编码风格 |
| L2 | 上下文 | CLAUDE.md、项目规则 |
| L3 | 策略 | 设计文档、计划 |
| L4 | 执行 | 任务、待办事项 |
| L5 | 对话 | 对话记录（过程） |
| L6 | 成果 | Git commit（结果） |
回到项目历史中的任意时间点，创建新分支——带有那个时刻的完整对话上下文，以及可选的"事后认知"（之后发生了什么）。
- macOS（Apple Silicon）
- 一个或多个支持的 AI 编程工具
从 Releases 下载最新的 DMG。
- 打开 DMG
- 将 Memora 拖入 Applications（应用程序）
- 启动 Memora——它会请求安装 hook 到你的 AI 工具
- 使用任何支持的工具开始编程——Memora 自动捕获
首次启动时，Memora 会：
- 请求安装 hook 到检测到的 AI 工具
- 导入现有的对话历史
- 展示界面交互式引导
Hook CLI → ~/.memora/events.jsonl + ~/.memora/raw/
↓
SQLite（索引）
↓
Flutter 桌面 UI
- 前端：Flutter（macOS 桌面）
- 后端：Rust（3 个 crate：memora-core、memora-ingest、memora-hook）
- 桥接：flutter_rust_bridge v2.12
- 数据：所有数据保留在本地。不会上传到任何地方。
Memora 100% 本地运行。你的对话、代码和数据永远不会离开你的机器。没有云同步、没有遥测、没有分析。
Memora 支持英文和中文。在设置中切换。
Leon Xu（云梦泽） — EastWorld ltd.
Memora 是专有软件。保留所有权利。