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 Anthropic SDK

Point the Anthropic SDK at Memtro and every Messages call gets your memory, connectors and routing.

Memtro serves the Anthropic Messages API at /v1/messages: system, content-block messages, max_tokens and tools with input_schema, answered in the same named-event stream the SDK already reads. Set ANTHROPIC_BASE_URL to platform.memtro.com and use a Memtro key as the API key; nothing else in your code changes. Model ids are Memtro's, auto, memtro or any provider/model you hold a key for, and the memtro field carries agent mode and the other options. count_tokens answers too: Anthropic's own tokeniser where the request is bound for an Anthropic model, a marked estimate otherwise.

What you get

  • Messages API with streaming, system prompts and client-declared tools, returned as tool_use blocks for you to run
  • Bearer token or x-api-key, and /v1/models with the fields Anthropic clients read
  • auto, memtro and provider/model ids, plus memtro options such as mode: agent

How to connect

  1. 01Create an API key on Dashboard → Profile → API keys
  2. 02Set ANTHROPIC_BASE_URL=https://platform.memtro.com and ANTHROPIC_API_KEY=sk-memtro-… (the SDK appends /v1/messages itself)

Example queries

  • › What changed on the Baltic account this week?
  • › {"memtro": {"mode": "agent"}} on a request runs a full investigation

Why it is better with memory

Through Anthropic SDK 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.