add-fal-model

What This Skill Does

The add-fal-model skill is a comprehensive developer tool designed to streamline the integration of new Fal AI models into the OpenClaw V3 API ecosystem. It provides an automated, structured framework for developers to register models that automatically become accessible across both the V3 API endpoints and the MCP (Model Context Protocol) interface. By utilizing this skill, you ensure consistent price calculation, schema validation, and execution logic for every new model added to your infrastructure.

Installation

To integrate this skill into your workflow, run the following command in your terminal: clawhub install openclaw/skills/skills/hexiaochun/add-fal-model Ensure your development environment is connected to the project source repository at openclaw/skills.

Use Cases

• Integrating new generative AI models (image, video, audio) from fal.ai into your proprietary V3 backend.
• Standardizing complex pricing models across diverse AI services, including fixed-cost, per-second billing, and duration-based tiers.
• Maintaining model registries in translate_api/app/api/v3/transports/mcp/model_registry.py without manual errors.
• Onboarding developers to the V3 architecture by enforcing strict input/output media type validation and documentation requirements.

Example Prompts

1. "I need to add the new Flux.1-schnell model from fal.ai to our V3 API. Here is the documentation link: https://fal.ai/models/fal-ai/flux/schnell. Please analyze the billing and help me configure the executor."
2. "Help me update the model registry for the latest video-to-audio model. It charges per second. What steps should I take in the executor file?"
3. "Add a new model to our V3 system. It uses duration-based pricing for video generation. I have the API schema here: [paste schema]."

Tips & Limitations

• Pricing Accuracy: Always verify the billing method directly on the fal.ai Playground page before implementing the executor code. Mistakes here lead to revenue discrepancies.
• Schema Validation: Always define your input/output media types clearly. V3 systems rely on these tags for proper route handling in the TaskServiceV3 architecture.
• Unified Architecture: Remember that the executor handles the heavy lifting; ensure that the SUPPORTED_CHANNELS constant is updated to include your new model, or it will remain inaccessible to the MCP interface.
• Testing: Always test both the REST API endpoint and the MCP submit_task function after registration to ensure complete integration path verification.