Astrolabe
Policy-driven OpenAI-compatible routing proxy for OpenClaw that picks the lowest-cost model, adds safety gates, and can escalate once.
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for Astrolabe, derived from its capabilities.
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology — every score is re-derived by the same automated method as an agent's public evidence changes.
Overview
Astrolabe is an open-source, policy-driven, OpenAI-compatible routing proxy built for OpenClaw. It sits between an OpenAI-compatible client (like OpenClaw) and OpenRouter, classifies each request into policy categories and complexity, applies safety gates, and routes the request to a cost-effective model with fallback chains for provider/model outages. For non-stream responses it can run a confidence self-check and optionally escalate once to a stronger model. Astrolabe is intentionally small and stateless (no database or UI required), returns routing metadata via response headers, and includes a health endpoint and structured logs for operational visibility.
Key features and capabilities
- openai-compatible proxy
- model routing
- cost optimization
- openrouter
- fallback chains
- safety gate
- self-check escalation
- routing metadata headers
- stateless server
- nodejs
Use cases
- Reducing OpenClaw inference costs by routing routine traffic to cheaper models and escalating only when needed.
- Adding a safety gate and policy-based routing layer in front of OpenRouter without changing client protocols.
- Improving reliability with fallback chains when a chosen model/provider is unavailable.
- Getting routing transparency via response metadata headers and structured logs for debugging and optimization.