MCP Server

Maintained

com.mcparmory/replicate MCP Server

com.mcparmory/replicate

Run AI models and manage predictions via Replicate's cloud API

What is the com.mcparmory/replicate MCP server?

The Replicate MCP server connects AI agents to Replicate's API for running machine learning models, creating deployments, and managing predictions. It exposes Replicate's full API surface as MCP tools, enabling programmatic access to thousands of open-source and commercial AI models.

This server lets you leverage Replicate's hosted AI models directly from Claude or Cursor. Run image generation, text processing, audio synthesis, and other ML tasks without managing infrastructure. Useful for agents that need to generate content, process data, or orchestrate multi-step AI workflows.

How to install com.mcparmory/replicate

Copy-paste configuration for popular MCP clients.

transport: stdio

Config generated by PluginBench — verify against the source before use.

~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "replicate": {
      "command": "uvx",
      "args": [
        "mcparmory-replicate"
      ]
    }
  }
}

Tools & capabilities

Tools this server exposes to the agent.

Run predictions — Execute AI models on Replicate with custom inputs and retrieve results
Create deployments — Set up persistent model deployments for repeated inference
Manage predictions — List, retrieve status, and manage prediction jobs

Use cases

Generate images or text using open-source models like Stable Diffusion or Llama
Automate batch processing of media files through ML pipelines
Build AI-powered agents that call multiple models in sequence
Create deployments for frequently-used models to reduce latency
Monitor and manage long-running prediction jobs

com.mcparmory/replicate MCP server FAQ

What is the Replicate MCP server?

It's an MCP server that exposes Replicate's API as tools, allowing Claude or Cursor to run AI models, create deployments, and manage predictions without leaving your editor.

Is it free to use?

The server itself is free (MIT license), but running models on Replicate incurs API costs based on compute usage. See Replicate's pricing page for details.

How do I install it in Cursor or Claude?

Add the server to your MCP client config with the command `uvx mcparmory-replicate` and set your Replicate API token in the `BEARER_TOKEN` environment variable.

What authentication is required?

You need a Replicate API token, which you can generate from your Replicate account dashboard.

What models can I run?

Any model available on Replicate's platform, including Stable Diffusion, Llama, Mistral, SDXL, and hundreds of community-contributed models.

Can I use this for batch processing?

Yes, the server supports creating predictions and managing them asynchronously, making it suitable for batch workflows.

README (reference)

Source of truth, from the repository.

<img src="https://wjxawmrpsfuivlicnepc.supabase.co/storage/v1/object/public/newsletter/logo-armory-no-background.png" alt="MCP Armory" height="64"> <h1 align="center">MCP Armory Registry</h1> Production-ready MCP servers for popular APIs, generated by <a href="https://mcpblacksmith.com">MCP Blacksmith</a>. <a href="https://github.com/mcparmory/registry/stargazers"><img src="https://img.shields.io/github/stars/mcparmory/registry?style=flat&color=gold" alt="GitHub Stars"></a> <a href="https://github.com/mcparmory/registry/blob/main/servers/github/LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue" alt="License"></a> <img src="https://img.shields.io/badge/servers-79-8A2BE2" alt="Servers"> <img src="https://img.shields.io/badge/python-3.11+-3776AB?logo=python&logoColor=white" alt="Python 3.11+"> <a href="#servers">Servers</a> · <a href="#install--usage">Install & Usage</a> · <a href="#features">Features</a> · <a href="#how-servers-are-generated">How it works</a> · <a href="#license">License</a>

MCP servers that connect AI agents to real-world APIs. Each server is a standalone Python package — generated from the upstream OpenAPI spec using MCP Blacksmith and tested against the live API before release. Hosted on MCP Armory or run independently.

Servers

🧠 AI & Machine Learning ElevenLabs · Linkup · Parallel · Perplexity AI · Ragie · Replicate

📊 Analytics Ahrefs · Datadog · Google Analytics · Google Search Console · Linkly · Mixpanel · PostHog

⛓️ Blockchain & Crypto Alchemy NFT

💼 CRM & Sales Apollo

💬 Communication Atlassian Confluence · Google Gmail · MailerSend · Mailtrap · Postmark · Resend · Slack

🗄️ Data & Databases Enigma · Pinecone

🛠️ Developer Tools AgentQL · Apify · Bitbucket · Bright Data · Browserbase · BuiltWith · Canvas · CircleCI · Codacy · Contentful Management · ConvertAPI · E2B · Firecrawl · GitHub · GitLab · LaunchDarkly · Postman · Scrapingant · Semrush AppCenter · Sentry

📄 Documents & Printing PDF.co

🛒 E-commerce Shopify Admin

💰 Finance Alpha Vantage · Polygon · Ramp

👥 HR & Recruiting BambooHR

☁️ Infrastructure Globalping · Grafana · PagerDuty · Rootly · Runpod

⚖️ Legal & Compliance NetLicensing

🗺️ Maps & Location Google Maps Platform · IP2Location.io IP Geolocation · OpenCage Geocoding

📣 Marketing Customer.io Journeys Track · Klaviyo

🎬 Media & Entertainment Canva · Perigon

✅ Productivity Airtable · Asana · Atlassian Jira · ClickUp · Figma · Google Sheets · Miro · Notion · Outline · Shortcut

🔍 Search Algolia Search · Google Search

📦 Storage Box · Files.com · Google Drive

📌 Other Close

More servers are being added continuously.

Install & Usage

Every server is a standalone PyPI package (mcparmory-<server-name>). Each server's own README.md has full credential setup and configuration instructions.

# With uvx (no install needed)
uvx mcparmory-github

# With pip
pip install mcparmory-github
mcparmory-github

MCP client configuration (Claude Desktop, Cursor, Codex, Claude Code):

{
  "mcpServers": {
    "github": {
      "command": "uvx",
      "args": ["mcparmory-github"],
      "env": {
        "BEARER_TOKEN": "ghp_your_token_here"
      }
    }
  }
}

Features

Every server in this registry provides:

Full API coverage — all endpoints from the OpenAPI spec are exposed as MCP tools
Authentication — API key, Bearer token, Basic auth, OAuth2, JWT, OpenID Connect, and mutual TLS
Multi-auth — servers that support multiple auth methods let you configure whichever you prefer
Pydantic validation — request parameters are validated before sending to the API
Resilience — configurable retries with exponential backoff, connection pooling, and timeouts
Response sanitization — optional redaction of sensitive fields from API responses

How servers are generated

Servers are generated by MCP Blacksmith, an automated pipeline that:

Parses an OpenAPI specification
Transforms operations through LLM-driven passes (naming, parameter filtering, enhancement, classification)
Generates a complete MCP server with auth, models, validation, and configuration
Tests every tool against the live API using an autonomous agent
Deploys passing servers to this registry

The generator targets Python 3.11+ and uses FastMCP as the MCP framework.

Contributing

Servers are auto-generated — don't edit server code directly. If you find an issue, open a GitHub Issue describing:

Broken or deprecated endpoints
Missing or incorrect parameters
Authentication problems
Incorrect request/response models

We'll re-generate the affected server with the fix. To request a new server, open an issue with a link to the API's OpenAPI spec or documentation.

License

Each generated server is released under the MIT License. The underlying APIs have their own terms of service — see individual server READMEs for details.

<a href="https://mcparmory.com">MCP Armory</a> · <a href="https://mcpblacksmith.com">MCP Blacksmith</a> · <a href="https://docs.mcpblacksmith.com/quickstart">Docs</a>

Related MCP servers

com.mcparmory/resend

Maintained

Send, track, and manage transactional emails at scale with Resend API

Python

View repository →

com.mcparmory/rootly

Maintained

Manage incident alerts, events, and workflows in Rootly with AI-powered automations

Python

View repository →

com.mcparmory/runpod

Maintained

Launch, scale, and manage GPU pods and serverless endpoints across regions with Runpod

Python

View repository →

com.mcparmory/scrapingant

Maintained

Web scraping with JavaScript rendering and CAPTCHA handling for AI agents

Python

View repository →

com.mcparmory/semrush-appcenter

Maintained

Manage Semrush app subscriptions, users, and send event notifications via MCP

Python

View repository →

com.mcparmory/sentry

Maintained

Track errors, manage performance alerts, and configure dashboards and monitors with Sentry via MCP.

Python

View repository →