Replicate Model Classifier

An HTTP API to classify Replicate models into Hugging Face tasks using a language model.

Powered by:

• Cloudflare Workers for hosting the HTTP API
• Cloudflare KV for caching
• Hono for authoring the HTTP API
• Anthropic Claude 3.7 Sonnet for model classification
• Replicate API for model metadata
• Hugging Face Tasks for model task metadata

Repository: https://github.com/zeke/replicate-model-classifier

API

Base URL: https://replicate-model-classifier.ziki.workers.dev/

Classify a model

GET /api/models/:owner/:model

Returns a JSON object with the model classification:

{
    "model": "salesforce/blip",
    "classification": {
        "summary": "Generate image captions and answer questions about images",
        "inputTypes": ["image", "text"],
        "outputTypes": ["text"],
        "task": "visual-question-answering",
        "taskSummary": "Visual Question Answering is the task of answering open-ended questions based on an image. They output natural language responses to natural language questions."
    }
}

Examples

• /api/models/bytedance/sdxl-lightning-4step
• /api/models/meta/meta-llama-3-8b-instruct
• /api/models/black-forest-labs/flux-schnell
• /api/models/salesforce/blip
• /api/models/meta/meta-llama-3-70b-instruct
• /api/models/stability-ai/stable-diffusion
• /api/models/abiruyt/text-extract-ocr
• /api/models/tencentarc/gfpgan
• /api/models/andreasjansson/clip-features
• /api/models/stability-ai/sdxl

View the prompt

To get a pretty-printed view of the prompt that was used to classify the model, add the prompt query parameter:

GET /api/models/:owner/:model?prompt=1

Examples

• /api/models/wavespeedai/wan-2.1-i2v-480p?prompt=1
• /api/models/meta/meta-llama-3-8b-instruct?prompt=1
• /api/models/black-forest-labs/flux-schnell?prompt=1

View everything

It can be helpful to see all the data that goes into the model classification. You can see all the data by adding the debug query parameter:

GET /api/models/:owner/:model?debug=1

Examples:

• /api/models/wavespeedai/wan-2.1-i2v-480p?debug=1
• /api/models/meta/meta-llama-3-8b-instruct?debug=1
• /api/models/black-forest-labs/flux-schnell?debug=1

Bust the cache

Responses are cached forever by default. To bust the cache for a specific model, use the force query parameter:

GET /api/models/:owner/:model?force=1

View Hugging Face task data

GET /api/tasks

See /api/tasks

View Hugging Face task names

GET /api/taskNames

See /api/taskNames

View all cached classifications

GET /api/classifications

Returns a JSON object containing all cached model classifications. Each key is the model identifier (owner/modelName) and the value is the classification data.

You can filter the results by task type using the task query parameter:

GET /api/classifications?task=text-generation

Examples:

• All text generation models
• All image-to-image models
• All text-to-image models
• All visual question answering models
• All image classification models

For a list of all available task types, see /api/taskNames.

Example response:

{
  "salesforce/blip": {
    "summary": "Generate image captions and answer questions about images",
    "inputTypes": ["image", "text"],
    "outputTypes": ["text"],
    "task": "visual-question-answering",
    "taskSummary": "Visual Question Answering is the task of answering open-ended questions based on an image. They output natural language responses to natural language questions."
  },
  "meta/meta-llama-3-8b-instruct": {
    "summary": "A large language model for text generation and instruction following",
    "inputTypes": ["text"],
    "outputTypes": ["text"],
    "task": "text-generation",
    "taskSummary": "Text generation is the task of generating text that is coherent and contextually relevant."
  }
}

See /api/classifications