NewCost control: a token allowance a month for each model choice — when one runs out, the cheaper one answers.See how it works
memtro
Start for free
Client

Memtro for goose

Point goose's OpenAI provider at Memtro, and add Memtro as a remote extension for its tools.

goose is Block's open-source local agent. Its OpenAI provider takes any compatible host, so OPENAI_HOST=https://platform.memtro.com makes every goose session run on Memtro's routing with your memory behind it, and a Streamable HTTP extension gives goose the memory and connector tools directly. Use either, or both.

What you get

  • OpenAI provider on a custom host: OPENAI_HOST https://platform.memtro.com, OPENAI_BASE_PATH v1/chat/completions, model auto
  • Remote extension over Streamable HTTP with an Authorization header, from the CLI, the desktop app or config.yaml
  • Recipes and sessions that read your systems and remember what they found

How to connect

  1. 01goose configure → Configure Providers → OpenAI: API key sk-memtro-…, OPENAI_HOST https://platform.memtro.com, OPENAI_BASE_PATH v1/chat/completions, model auto (or export OPENAI_API_KEY and OPENAI_HOST)
  2. 02goose configure → Add Extension → Remote Extension (Streamable HTTP): name memtro, URL https://platform.memtro.com/api/mcp, custom header Authorization: Bearer sk-memtro-…
  3. 03Or in config.yaml: extensions: memtro: { type: streamable_http, uri: https://platform.memtro.com/api/mcp, headers: { Authorization: Bearer sk-memtro-… }, enabled: true }

Example queries

  • › What did we promise Trust Payments last week?
  • › Fix the webhook retry, and check Sentry through memtro first

Why it is better with memory

Through goose you get the same memory, connector tools, routing and jobs as every other client, so what one person learns in one tool is available to the whole team in all of them.