Open Registry & Intelligence Platform for AI Infrastructure
Discover, compare and monitor LLM models, inference providers, MCP servers and agent skills using open data and real-time telemetry.
Track latency, uptime, pricing, rate limits, capabilities, and provider mappings — in one place.
| Models | Providers | Mappings | Skills | MCP Servers |
|---|---|---|---|---|
| 102 | 49 | 159 | 80+ | 50+ |
- Models — canonical registry of LLM models: capabilities, context windows, modalities, licensing, open-weight status
- Providers — 49 inference providers with API base URLs, auth types, regions, free tiers, trial credits
- Mappings — per-provider rate limits, pricing, and model availability
- Telemetry — real-time latency (TTFT p50/p95/p99) and uptime monitoring across providers
- MCP Servers — registry of 50+ Model Context Protocol servers
- Skills — 80+ ready-to-use AI agent prompts and workflows for Claude Code, Cursor, Kiro, Copilot, Windsurf and more
- Insights — data-driven analysis and benchmarks based on live telemetry
- CLI — command-line interface for querying the registry
openmodels/ → public YAML registry (models, providers, mappings, schemas)
skills/ → community AI agent skills (github.com/openmodelsrun/skills)
cms/ → Payload CMS for Insights content (cms.openmodels.run)
docs/ → architecture, roadmap, contributing guides (docs.openmodels.run)
platform/ → intelligence platform — Next.js + NestJS + Python workers
→ openmodels.run — search, compare, telemetry, skills, MCP servers, insights
- Registry (models, providers, mappings)
- Telemetry (Anthropic, Groq, AWS, DeepSeek, Alibaba, Google AI)
- Skills
- MCP Servers registry
- Insights
- CLI
- Public API
- SDK (
@openmodels/sdk) - User accounts + Pro plan
- i18n (Chinese, Russian, Spanish)
The registry is community-maintained. Adding a model, provider, or MCP server is a single YAML file.
