{
  "markdown": "# llm-oracle\n\n> Continuously-updated catalog + query API for LLM provider availability and pricing.\n\nBuilt by [weiseer](https://weiseer.com).\n\n## What it does\n\nGiven a question like *\"for this prompt of N input tokens plus M expected output tokens, which currently-available model offers the lowest cost?\"* the oracle answers in milliseconds, with the source URL it used for the pricing claim.\n\nUse it as:\n- an MCP server for AI agents that need to make routing decisions at inference time\n- an HTTP JSON API for any tool that wants up-to-date LLM pricing/availability\n- a raw `catalog.json` you can audit, vendor, or contribute fixes to\n\n## Why\n\nLLM pricing changes monthly. Every multi-model app rebuilds the same lookup table. We maintain it so you don't.\n\n- **Cited**: every pricing row links to the provider's official documentation\n- **Versioned**: `as_of` timestamp on every response so you know how fresh the data is\n- **Open**: catalog is public JSON, MIT licensed; client/server code is Apache-2.0\n- **Cross-checked**: source data verified against the open-source [LiteLLM price file](https://github.com/BerriAI/litellm) where overlap exists\n\n## Coverage (v0)\n\n5 providers, ~10 models. Growing as we observe demand.\n\n- Anthropic: Claude Opus 4.7, Sonnet 4.6, Haiku 4.5\n- OpenAI: GPT-5, GPT-4o\n- Google: Gemini 2.5 Pro, Gemini 2.5 Flash\n- DeepSeek: Reasoner, Chat\n- Mistral: Large\n\n## Quickstart\n\n### As an MCP server (Claude Desktop, Cursor, Continue, Cline, etc.)\n\n```bash\ngit clone https://github.com/weiseer/llm-oracle\ncd llm-oracle\n```\n\nAdd to your MCP client config:\n\n```json\n{\n  \"mcpServers\": {\n    \"llm-oracle\": {\n      \"command\": \"python\",\n      \"args\": [\"/absolute/path/to/llm-oracle/mcp_server.py\"]\n    }\n  }\n}\n```\n\nThe agent now has tools:\n\n- `list_models(provider?, capability?)` → list models, optionally filtered\n- `get_model(model_id)` → full record for one model\n- `find_cheapest(input_tokens, output_tokens, required_capabilities?)` → ranked by estimated cost\n- `compare_models(model_ids[], input_tokens?, output_tokens?)` → side-by-side\n- `check_availability(model_id)` → current status with cited source\n\n### As an HTTP API\n\n```bash\n# Free tier: 1,000 calls/day per IP, no key\ncurl https://oracle.weiseer.com/catalog.json\ncurl 'https://oracle.weiseer.com/cheapest?input_tokens=2000&output_tokens=500'\ncurl https://oracle.weiseer.com/models/claude-sonnet-4-6\n```\n\n### As raw JSON\n\n```bash\ncurl https://oracle.weiseer.com/catalog.json > my-local-catalog.json\n```\n\n## Pricing\n\n| Tier | Calls | Cost |\n|------|-------|------|\n| Free | 1,000/day | $0 |\n| Pro | 100,000/month | $5 USDC/mo |\n| Scale | 1,000,000/month | $20 USDC/mo |\n| Raw catalog | unlimited | $0 (always free) |\n\nPaid tiers settle in USDC. Email `wei@weiseer.com` to subscribe; we issue an API key and confirm receipt. (We'll automate this when the first three paid customers exist.)\n\n## Schema\n\nSee `catalog.json` for the canonical schema. Per-model fields include `model_id`, `provider`, `family`, `context_window`, `max_output_tokens`, `input_price`, `output_price`, `cached_input_price`, `capabilities`, `availability_status`, source URLs, and timestamps.\n\n## Update cadence\n\n- Pricing: daily cron + manual when a price change is publicly announced\n- Availability: 5-minute poll of provider status pages where available; daily smoke-test otherwise\n\n## Errors and disagreements\n\nFound a stale price? An incorrect capability flag? A missing model? [Open an issue](https://github.com/weiseer/llm-oracle/issues) — please include the source URL we should be tracking.\n\n## Telemetry\n\nThe hosted service logs each request as a single event (caller, query summary, latency, paid status) to weiseer's append-only audit ledger. The MCP server you self-host logs nothing.\n\n## Building on this\n\nThe catalog format is MIT licensed. Vendor it, mirror it, build on top of it. We ask only that you cite `weiseer/llm-oracle` somewhere visible if you redistribute the data.\n\nThe server and MCP code are Apache-2.0.\n\n## Roadmap (probe-pulled, not pre-planned)\n\nThis is **P-001 in weiseer's strategic backlog** — a market sensor for organism's architecture v2. We expand it where users ask, kill it if 4 weeks of public availability brings under 100 calls. Don't expect a polished commercial product; expect honest experimentation with daily updates.\n\nWhat we might add (in observed-demand order):\n- More providers (Cohere, xAI, AI21, Together, Replicate hosted, Bedrock, Azure pricing variants)\n- x402 micropayments for high-volume agent buyers (deferred — see [arXiv 2605.11781](https://arxiv.org/abs/2605.11781) on protocol security)\n- Historical pricing time-series\n- SLA + uptime aggregates per provider\n- Per-region availability where providers vary\n- Self-hosted model cost estimators (with caveats)\n\n## Contact\n\n`wei@weiseer.com` · [github.com/weiseer](https://github.com/weiseer) · [weiseer.com](https://weiseer.com)\n",
  "bytes": 4901,
  "sha": "54cf9284f9ced6ebf1660813edee0ce83e282d2910c5b05c08d23a95c0e19127",
  "repo_slug": "weiseer/llm-oracle",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_weiseer_llm_oracle_89e5d723/readme"
}