A unified gateway providing access to over 70 leading LLMs through a single OpenAI-compatible API.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install aisa-llm-router-skill
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 aisa-llm-router-skill 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 LLM Router?
The OpenClaw LLM Router is a sophisticated abstraction layer designed for developers and autonomous agents who require access to a diverse range of Large Language Models. By consolidating providers like OpenAI, Anthropic, Google, and DeepSeek into a single interface, it eliminates the need to manage dozens of separate API keys and SDKs. This skill is a core part of the Openclaw Skills library, enabling technical users to switch between model families simply by changing a string in their configuration.
Beyond simple access, the router provides a standardized way to handle vision analysis, function calling, and streaming responses. It simplifies the development lifecycle by offering unified billing and cost tracking, ensuring that agents can scale across different model architectures without the overhead of fragmented infrastructure.
OpenClaw LLM Router Use Cases
- Reducing architectural complexity by replacing multiple provider-specific SDKs with a single OpenAI-compatible endpoint.
- Implementing cost-optimization logic that routes high-volume, simple queries to affordable models like Gemini Flash while reserving complex reasoning for GPT-4 or Claude 3.5.
- Building resilient AI applications that utilize automatic fallback strategies to maintain uptime when a specific provider experiences latency or outages.
- Conducting automated A/B testing and model comparison to determine which LLM performs best for specific domain-driven tasks.
How OpenClaw LLM Router Works
- The application sends a standardized request to the router's unified endpoint with the AISA API key.
- The router identifies the requested model (e.g., gpt-4o, claude-3-sonnet, or deepseek-v3) and maps it to the appropriate provider backend.
- The router handles any necessary protocol translation, such as converting native Gemini formats to OpenAI-compatible structures.
- The response is returned to the user in a consistent JSON format, including enriched metadata for usage tracking and precise cost calculation.
OpenClaw LLM Router Setup
To begin using this skill, you must first obtain an API key from the AIsa dashboard. Once acquired, configure your environment using the following command:
export AISA_API_KEY="your-key"
Because the Python client is designed for maximum compatibility, it uses the standard library and requires no external dependencies. You can verify your setup by running the built-in client script to query a model:
python3 scripts/llm_router_client.py chat --model gpt-4o --message "Hello from Openclaw Skills"
OpenClaw LLM Router Data Schema & Taxonomy
The router organizes interaction data using a standardized schema that mirrors the OpenAI Chat Completions API. This ensures that any tool compatible with Openclaw Skills can easily parse the responses.
| Key | Type | Description |
|---|---|---|
model |
string | The specific identifier for the LLM used for the request. |
choices |
array | Contains the generated message and finish reason. |
usage.prompt_tokens |
integer | Number of tokens processed in the input. |
usage.cost |
number | The calculated cost of the request in USD. |
usage.credits_remaining |
number | The updated balance of the unified account. |
OpenClaw LLM Router Advanced Features
- Multi-model comparison engine to evaluate responses from multiple providers for the same prompt simultaneously.
- Native vision and multimodal support for processing image URLs and base64 data across supported models.
- Integrated function calling support for structured data extraction and agentic tool use.
- Global streaming support (SSE) to enable real-time UI updates regardless of the underlying provider's native capabilities.
SKILL.md
Loading
Related Openclaw Skills
Model Failover Guard
An automated health monitoring and failover utility that ensures continuous AI agent uptime by switching providers during outages.
Ansible Infrastructure Automation
Automate server provisioning, configuration management, and application deployment using Infrastructure as Code.
TiDB Cloud Zero
A specialized tool for provisioning ephemeral, disposable TiDB Cloud Zero databases designed for automated agent workflows and rapid testing.
LocalClaws Meetup Coordinator
A privacy-first agent skill designed to manage local meetup logistics for hosts and attendees via the LocalClaws protocol.
OpenClaw Media Gen
A powerful multimedia generation skill that unifies Gemini image and Qwen Wan video creation through a single API key.
Agent Access Control
A sophisticated security and permission layer providing tiered access control, owner approval workflows, and automated stranger deflection for AI agents.
METADATA
Featured*
