RunAPI MCP Server
RunAPI MCP server: a catalog of 160 models for images, video, music and speech, pricing, task runs and balance
- 56 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Created tasks spend the paid RunAPI balance
- The API key sits in a config file on disk or in an environment variable
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/runapi-mcp-serverDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Created tasks spend the paid RunAPI balance
- The API key sits in a config file on disk or in an environment variable
This is third-party code. Review the repository files before installing.
What it does
RunAPI MCP Server connects Claude Code, Codex, Cursor and other clients to RunAPI, which collects models for image, video, music and speech generation, more than 160 options. The agent browses the catalog, reads a model's inputs and current price, creates media tasks, polls their status and checks the balance. The catalog, pricing and prompt search tools work without a key because the catalog is embedded in the package. Creating tasks and checking the balance need sign-in: the `login` tool opens a browser and in CI you set `RUNAPI_API_KEY`. There is a local stdio mode and a hosted address `https://mcp.runapi.ai/mcp` with eight tools and no `login`.
Who it is for. Developers and creators who run image, video and audio generation from an agent terminal and track cost.
Good fit when
- Before generating a clip you want to compare model prices and pick one that fits the input fields
- The agent creates an image task and waits for the result in the same dialogue
- A CI build must check the balance before batch generation
- You have a frame idea and need a ready prompt with an example image
Not a fit when
- Generation goes through another provider and RunAPI is not connected there
- You need local generation on your own GPU: the package is only a client and runs no backend
- All tasks are created by hand in the web interface and no agent takes part
Example request
Show the RunAPI video models, compare prices and create a task for a five-second clip of a sunset over the seaMCP
- Transport
- stdio, http
- Authentication
- key is optional
Environment variables
RUNAPI_API_KEYsecret- RunAPI API key for environments without a browser, replaces `login` sign-in.
Limitations
Creating tasks, status and balance work only after signing in to RunAPI, and tasks spend a paid balance. The package is a pure client and generates nothing locally. Sign-in data is saved to `~/.config/runapi/config.json`, the same file the `runapi login` command uses. The catalog is embedded at build time and prices in it are verified with the `check_pricing` tool. The hosted MCP has no `login` tool.
How to disable. Remove the server (`claude mcp remove runapi`), sign out of RunAPI and delete `~/.config/runapi/config.json`; revoke the API key in the dashboard.
FAQ
How do I install RunAPI MCP in Claude Code?
With `claude mcp add runapi -s user -- npx -y @runapi.ai/mcp`. The `-s project` flag writes the server into the repository `.mcp.json` for the whole team.
Is a key needed for the catalog and prices?
No, `list_models`, `get_model_info`, `check_pricing` and `search_prompts` work without sign-in.
How do I sign in on a server without a browser?
Set `RUNAPI_API_KEY` before starting the client, it replaces browser sign-in.
Is there a hosted version?
Yes, `https://mcp.runapi.ai/mcp` with OAuth or an API key as a bearer token, with no Node.js on the machine.
How do I avoid creating a task twice?
`create_task` needs an `idempotency_key` generated by the caller, so a repeat with the same key does not multiply tasks.
Discussion
No questions or reviews yet.
Related
MCP serversWaveSpeed MCP server: an agent runs image, video, audio and 3D models, reads each model's input schema and quotes the price first
An MCP server for generating images and video from the agent: several models, a large prompt library and parallel tasks
Hosted Magic Hour MCP server: the agent generates and edits video, images and audio through the service API
The Runway Dev plugin gives an agent an MCP server and seven skills to embed Runway video, image and audio generation
MCP server for image generation and editing: Gemini, OpenAI and Seedream with automatic prompt enhancement
A community MCP server for the Kling AI API: video, images, lip sync and virtual try-on. It uploads local files to a third-party cloud bucket
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| RunAPI MCP Server (this card) | yes | key is optional | 56 | |
| ai-media-generator | no | 274 | A single entry point for media models, while RunAPI has an embedded catalog with prices and an account balance. | |
| MCP Image Generator | no | key is optional | 190 | Images only, while RunAPI also covers video, music and speech. |
README in short
A package description, the hosted MCP, a quick start for Claude Code, a table of nine tools with the sign-in flag, example requests and scenarios for the catalog, prices, prompts and tasks.