Replicate Flux MCP Server
Flux-based image and SVG generation capabilities via Replicate API endpoints.
Data last scanned today · Reviewed today
Overview
Replicate Flux MCP allows AI assistants to generate high-quality raster images and vector graphics directly through the Replicate platform. By default, it integrates with the Flux Schnell model for image creation and Recraft v3 for SVG output. The server supports model customization via environment variables and allows per-call model overrides through a built-in allowlist. This design provides users with granular control over the generative models used during their workflow.
Our verdict
With a security score of 45/100, this server is currently classified as high risk. This rating reflects that the project is community-maintained rather than vendor-backed, and several critical areas such as authentication methods and read-only mode support remain unreviewed. Given the potential for unauthorized API usage if misconfigured, this tool is best suited for developers comfortable auditing community-sourced code before local deployment.
- Uses only three direct dependencies to minimize supply-chain exposure
- Provides built-in support for both raster and SVG generation
- Supports model overrides for flexible generative workflows
- Maintains active repository status with recent commit history
- Authentication mechanisms have not undergone formal security review
- Community-maintained status lacks official vendor support or auditing
- Read-only mode operations remain currently unverified
Setup
{
"mcpServers": {
"replicate-flux-mcp": {
"args": [
"-y",
"replicate-flux-mcp"
],
"command": "npx"
}
}
}Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.
- No read-only mode — running this grants full read/write access.
- No built-in authentication — anyone with access to this config can invoke every tool it exposes.
Tools
| Tool | Description | Risk |
|---|---|---|
| generate_image | Generates an image based on a text prompt using the configured image model (allowlist only). | medium |
| generate_multiple_images | Generates multiple images based on an array of prompts using the configured image model (allowlist only). | medium |
| generate_image_variants | Generates multiple variants of the same image from a single prompt using the configured image model (allowlist only). | medium |
| generate_svg | Generates SVG/vector output based on a text prompt using the configured SVG model (allowlist only). | medium |
| prediction_list | Retrieves a list of your recent predictions from Replicate. | low |
| get_prediction | Gets detailed information about a specific prediction. | low |
| run_model | Runs a whitelisted Replicate model with a raw input payload (useful for video or restoration models). | medium |
| run_replicate_model | Runs any model hosted on Replicate by its owner/name[:version] reference. | medium |
| get_model_schema | Fetches the OpenAPI input schema and description for a Replicate model so you can pass the right parameters to run_replicate_model. | low |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›What prerequisites do I need to use this server?
You need to obtain a Replicate API token from your Replicate account settings to authenticate with the service.
›Can I use different models for image generation?
Yes, you can override the default models using environment variables, or provide a specific model_id from the built-in allowlist during an individual tool call.
›Which models does this server use by default?
By default, it utilizes black-forest-labs/flux-schnell for raster image generation and recraft-ai/recraft-v3-svg for SVG output.
›Are there alternatives to running this server locally?
Yes, you can use the server as a hosted service via Smithery or Glama.ai, which avoids the need for manual local setup.
Changelog
- New
- New[Version 0.1.1] - 2025-03-17
2025-03-17
- Fix[Version 0.1.0] - 2025-03-16
2025-03-16
Badge
Maintain this server? Add the live badge to your README.