canonical: https://jentic.com/apis/creativeclaw.co/creativeclaw # CreativeClaw MCP API The CreativeClaw API generates video, 3D models, and audio for AI agents through a remote tool server that speaks JSON-RPC over a single streamable HTTP endpoint rather than a resource-oriented REST surface. Agents authenticate with OAuth 2.1, list the available tools, and invoke them. Generation tools such as video creation, HTML-to-video rendering, 3D model generation, transcription, audio isolation, and media merging run asynchronously and return a job to poll until it completes or fails. ## For AI agents Generate video, 3D models, and audio through a remote MCP server by calling its streamable HTTP endpoint, then poll the returned job until it finishes. Authenticated with OAuth 2.1 or a bearer JWT; discovery and health endpoints are public. ## Scope Does not handle media hosting, image editing, or payment processing. Use for generating video, 3D models, and audio through its MCP endpoint only. ## Capabilities - List the live generation tools and their schemas over the MCP endpoint - Generate a video from a prompt and render an HTML page to a video - Generate a 3D model from an input - Transcribe audio and isolate an audio track - Merge multiple media assets into one output - Poll an asynchronous generation job until it completes or fails ## Use cases ### Asynchronous Media Generation An agent lists the CreativeClaw tools over the MCP endpoint, invokes video, 3D, or audio generation with its inputs, and receives a job it polls until the result is ready. Because generation runs asynchronously, the agent calls the job-check tool rather than blocking, then uses the finished asset once the status reports completion. Example prompt: Call POST /mcp with a tools/call request to generate a video, then poll the returned job with the job-check tool until it completes ### Audio Processing for Agents An agent transcribes an audio file, isolates a track, or merges clips by invoking the matching CreativeClaw tool over the MCP endpoint. The tools accept the source media and return a job, so the agent can chain transcription and audio isolation into a larger content pipeline. Example prompt: Call POST /mcp with a tools/call request to transcribe an audio file, then poll the job until the transcript is ready ### AI Agent Integration via Jentic An AI agent reaches CreativeClaw through Jentic without managing the OAuth flow itself. It searches by intent, receives the matching operation and its input schema, and executes the call with the credential injected at run time. This lets an agent request a generated asset as one step in a task without handling the MCP client setup in code. Example prompt: Search Jentic for 'generate a video', load the CreativeClaw MCP operation schema, and execute it with Jentic-managed credentials ## Key endpoints | Method | Path | Description | | --- | --- | --- | | POST | `/mcp` | MCP Streamable HTTP endpoint for tool listing and calls | | GET | `/health` | MCP host health | | GET | `/.well-known/oauth-authorization-server` | OAuth authorization server metadata | | GET | `/.well-known/openid-configuration` | OpenID provider configuration | ## Key resources - **MCP endpoint** — List and invoke generation tools over JSON-RPC via /mcp - **Health** — Check MCP host health via /health - **OAuth discovery** — Read OAuth and OpenID configuration via the /.well-known endpoints ## AI readiness This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use. - **Score:** 69 / 100 - **Maturity:** AI-Aware - **Dimensions:** - Foundational Compliance: 100 / 100 - Developer Experience & Jentic Compatibility: 60 / 100 - AI-Readiness & Agent Experience: 48 / 100 - Agent Usability: 94 / 100 - Security: 83 / 100 - AI Discoverability: 75 / 100 - **View full report:** https://jentic.com/apis/creativeclaw.co/creativeclaw/scorecard - **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/ - **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview ### Score it yourself Every API in the directory is allowlisted, so you can re-score it with no key required. - **Score your own API:** https://jentic.com/scorecard.md - **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md ```sh npx @jentic/api-scorecard-cli score ``` ## Why Jentic - **Setup:** Wiring CreativeClaw by hand means running its OAuth 2.1 flow with PKCE, speaking JSON-RPC over its streamable HTTP endpoint, and polling each asynchronous generation job yourself. Through Jentic you install once, import CreativeClaw from the API Directory, store the credential once, and your agent calls it. - **Permission scoping:** CreativeClaw routes every tool through a single tool-call operation, with the tool and its inputs carried in the request body, so a rule bounds whether your agent may reach that operation rather than which tool it invokes. You choose to allow or withhold that operation, and the public health and discovery endpoints need no credential at all. - **Credential handling:** Your CreativeClaw credential is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'generate a video' or 'transcribe audio', and Jentic returns the matching CreativeClaw operation with its input schema so the agent calls the right endpoint without browsing the reference docs. ## Related APIs - **Replicate** — Run and host open generative models for image, video, and audio - **Runway** — Generative video models and editing tools - **ElevenLabs** — AI speech synthesis and voice generation ## FAQ ### What authentication does the CreativeClaw MCP API use? The MCP endpoint authenticates with OAuth 2.1 using the authorization code flow with PKCE, or a bearer JWT obtained from the token endpoint, per its OpenAPI spec. The health and OAuth discovery endpoints are public. Through Jentic the credential is stored encrypted by your own Jentic One instance and injected at execution time, so it never reaches the agent. ### Can I generate a video with the CreativeClaw MCP API? Yes. You call the video-generation tool over the MCP endpoint, which returns a job that runs asynchronously, then poll the job-check tool until its status reports completion. The same pattern covers HTML-to-video rendering, 3D model generation, transcription, audio isolation, and media merging. ### What are the rate limits for the CreativeClaw MCP API? The OpenAPI spec does not publish specific rate-limit numbers; check https://creativeclaw.co/developers for current details. Because generation runs as an asynchronous job, an agent invokes a tool and then polls its job rather than issuing a burst of requests. ### Is there a CreativeClaw MCP server? Yes, CreativeClaw is itself a remote MCP server reached over OAuth 2.1. You can also connect it directly through Jentic: your credential is injected at call time by your self-hosted instance instead of living in an MCP client config, your rules decide whether the agent may reach the endpoint, and every call is logged. Connecting the next API is then an import rather than another install and auth step. ### Can I limit what my agent is allowed to do with the CreativeClaw MCP API? Yes. CreativeClaw routes every tool through one /mcp operation with the tool and its inputs in the request body, so a rule allows or withholds that operation as a whole rather than scoping individual tools. Because every call runs through your own Jentic One instance, each invocation is logged for you to review. ### How do I generate media with the CreativeClaw MCP API through Jentic? Search Jentic for 'generate a video', and it returns the CreativeClaw MCP operation with its input schema. Load the schema and execute POST /mcp with a tools/call request, with the credential injected at run time. To run it on your own infrastructure, install Jentic One from its GitHub repo.