A specialized skill for configuring and validating custom AI model providers and local LLMs within the OpenClaw ecosystem.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install openclaw-model-provider
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install openclaw-model-provider using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is OpenClaw Model Provider Skill?
The openclaw-model-provider skill is an essential utility for developers looking to expand the capabilities of their AI agents beyond standard defaults. It facilitates the seamless integration of custom OpenAI-compatible APIs, local LLMs like Ollama, and specialized third-party services. By utilizing this skill, users can manage complex provider configurations, secure API keys via environment variables, and ensure high reliability through built-in validation tools.
As a core component of Openclaw Skills, this tool bridges the gap between various inference backends and the OpenClaw agentic framework. It streamlines the process of defining model parameters such as context windows, token limits, and reasoning capabilities, ensuring that your agents always have access to the most effective models for their specific tasks.
OpenClaw Model Provider Skill Use Cases
- Integrating local LLMs (e.g., Llama 3 via Ollama) for privacy-focused development.
- Adding third-party OpenAI-compatible proxies or specialized providers like Moonshot AI and SiliconFlow.
- Configuring fallback models and multi-key rotation for production-grade reliability.
- Setting up custom reasoning models and vision-capable endpoints for advanced agent workflows.
How OpenClaw Model Provider Skill Works
- Information Gathering: The skill identifies the provider type (Built-in vs. Custom) and collects essential metadata including Base URL, API Key, and Model IDs.
- Configuration Generation: It generates a structured JSON snippet that merges provider definitions with agent default settings.
- Application: The user is guided to apply the configuration to the local openclaw.json file using CLI tools.
- Validation: The skill triggers automated validation commands to ensure syntax correctness and connection stability.
- Deployment: Once validated, the new models are listed and ready for immediate use by OpenClaw agents.
OpenClaw Model Provider Skill Setup
To begin using this skill within the Openclaw Skills framework, ensure your environment is ready:
# Backup your existing configuration
cp ~/.openclaw/openclaw.json ~/.openclaw/openclaw.json.bak
# Open the configuration editor
openclaw config edit
# Set your API Key in your shell profile (e.g., .zshrc or .bashrc)
export PROVIDER_API_KEY="your-api-key-here"
OpenClaw Model Provider Skill Data Schema & Taxonomy
The skill organizes provider data within the models and agents nodes of the OpenClaw configuration. Below is the primary structure:
| Key | Description | Example |
|---|---|---|
models.providers |
Map of provider definitions and their endpoints. | "ollama-local": { ... } |
baseUrl |
The API endpoint for the provider. | [https://api.openai.com/v1](https://api.openai.com/v1) |
api |
The interface type (protocol). | openai-completions |
models[] |
Array of supported model objects including contextWindow and maxTokens. |
[{ "id": "gpt-4" }] |
agents.defaults.model |
Defines the primary model used by agents globally. | "provider/model-id" |
OpenClaw Model Provider Skill Advanced Features
- Multi-Key Rotation: Configure multiple API keys to enable automatic failover and bypass rate limits.
- Environment Variable Mapping: Use
${ENV_VAR}syntax to keep sensitive credentials out of plain-text config files. - Parameter Overrides: Fine-tune specific model behaviors (temperature, maxTokens, fastMode) at the agent level.
- Verbosity Control: Use the
--verboseflag during validation and testing to debug connection handshakes and API responses.
SKILL.md
Loading
Related Openclaw Skills
Feynman Technique Explainer
A research-backed learning tool that deconstructs complex subjects using the Feynman Technique to ensure deep, jargon-free understanding.
Urea Fertilizer Reference
A comprehensive technical reference for urea fertilizer management, covering nitrogen chemistry, application methods, and volatilization loss prevention.
Freeport
A comprehensive reference tool for managing free-trade zone regulations, duty deferral strategies, and bonded logistics operations.
Community Group Buying Supply Chain Management
A comprehensive end-to-end toolkit for managing supplier development, procurement negotiations, and logistics within the community group buying ecosystem.
MindMatch
MindMatch is an AI-driven human compatibility engine that uses psychological profiling rather than surface-level interests to find your ideal partners.
Web PPTX Generator
A professional single-page HTML presentation generator that creates beautiful, mobile-friendly slides with premium themes and no external API dependencies.
Featured*
