Skip to main content
Oimo’s MCP server lets coding agents and chat clients look up current list prices of AI model APIs and GPU compute, and which providers sell a model, for example Claude Opus 5.5 at Anthropic, Amazon Bedrock, Google Vertex AI and Microsoft Foundry. It needs no account and no key. The server URL is https://mcp.oimo.com/public (Streamable HTTP).

Connect a client

Claude Code:
Cursor: add it to ~/.cursor/mcp.json (or .cursor/mcp.json in a project):
mcp.json
VS Code: add it to .vscode/mcp.json:
mcp.json
Claude (web and desktop): Settings → Connectors → Add custom connector, then paste the URL. Any other client that speaks MCP over Streamable HTTP connects to the same URL.

Tools

Every tool only reads. Answers carry pricesAsOf, when Oimo last confirmed the prices, and a link to the same item in the app. Questions it answers well:
  • Which providers sell Claude Opus 5.5, and what does each charge per million input and output tokens?
  • What does Claude Sonnet 5.5 cost on Amazon Bedrock?
  • What does an H100 GPU hour cost at each provider?
  • Which providers offer a batch price for Claude Opus 5.5, and what is it at each?
  • Which image generation models have prices?

What it does not do

  • It serves current list prices as each provider publishes them: no negotiated discounts, credits or taxes, and no ranking or claim that two offers are equivalent.
  • get_prices shows each provider’s standard price, in its default region with the range across regions; compare_prices lists every plan and region. Each answer is about one item: it does not list a provider’s whole price list.
  • compare_prices returns at most 50 offers, shared across the providers so each one appears, with no paging: narrow it to one provider or one plan to see the rest.
  • Each address may send 30 requests at once, then 1 a second; past that it gets 429 with a Retry-After header.
  • Price history and change alerts need an account: the price history API and webhooks.