Replicate

Replicate MCP Connector for Claude

A+

Automate machine learning workflows via Replicate — run models, manage predictions, and search for AI assets directly from any AI agent.

20 tools Official Updated Oct 1, 2026 Official Vinkius Partner

Connect your Replicate account to any AI agent to run powerful machine learning models through simple commands.

What you can do

  • Run Predictions — Execute any model version with custom inputs and handle webhooks for completion
  • Model Discovery — Search for public models or list specific versions and their OpenAPI schemas
  • Management — Create, update, and manage your own models and hardware configurations
  • Monitoring — Track prediction status, logs, and outputs in real-time
  • Deployment — Manage deployments and training sessions for custom fine-tuned models

How it works

  1. Subscribe to this server
  2. Enter your Replicate API Token
  3. Start generating images, text, or audio from Claude, Cursor, or any MCP-compatible client

Who is this for?

  • AI Engineers — quickly test model versions and inputs without writing boilerplate code
  • Creative Technologists — generate media assets directly within their workflow
  • Data Scientists — monitor training jobs and prediction outputs from a unified interface
machine-learningai-modelsgpu-computinginference-apistable-diffusion

20 tools expose this connector's capabilities to your AI agent.

get_account

Get details about the authenticated user or organization

list_predictions

List recent predictions

search_models

Provide the search query. Search for public models on Replicate

create_prediction

Create a new prediction to run a model

create_deployment_prediction

Provide the deployment owner, name, and input parameters. Create a prediction using a dedicated deployment

create_model

Provide the owner, name, and required details. Create a new model on Replicate

create_training

Provide base model details and input data. Create a training to fine-tune an existing model

get_model_version

Provide owner, name, and version ID. Get details of a specific model version, including its OpenAPI schema

get_training

Provide a valid training ID. Get the status of a training

get_webhook_secret

Get the default webhook secret key for verifying signatures

list_collections

Do not use this tool to search for specific models. List curated collections of models

list_hardware

List available hardware SKUs and their descriptions

list_model_versions

Provide the owner and model name. List all versions of a model

cancel_prediction

Provide the prediction ID. Cancel a running prediction

create_deployment

Specify the deployment name and model. Create a new private deployment with custom autoscaling

delete_model_version

Provide the owner, model name, and version ID. Delete a specific model version

get_collection

Use the collection’s unique slug. Get details of a specific model collection

get_model

Provide both the owner and name. Get details of a specific model

get_prediction

Provide the prediction ID. Get the status and output of a prediction

update_model

Specify the owner and model name. Update metadata for an existing model

See how to talk to your AI agent using Replicate.

Search for public models related to 'Stable Diffusion' on Replicate.

I found several models. The most popular is 'stability-ai/stable-diffusion' (Version: 328bd9...). Would you like to see the input schema for this model?

Run a prediction for model version 5c51d4... with the prompt 'A futuristic city'.

Prediction created (ID: pred_123abc). It is currently 'starting'. I will monitor it for you or you can ask me for the status in a moment.

List my most recent predictions.

Fetching your history... You have 3 recent predictions. The latest one (pred_123abc) succeeded and generated an image URL. Would you like to see the results?

Use the `get_prediction` tool with your Prediction ID. It will return the current status (starting, processing, succeeded, or failed) along with the output URLs or data once completed.

Related Connectors