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 Codex

Run Codex on Memtro as its model, give Codex Memtro's tools over MCP, or run Memtro on your ChatGPT plan.

Three fits, and they combine. Codex speaks the Responses API to a custom provider, which Memtro serves at /v1/responses, so every Codex turn can run with your organisation's memory, connectors and routing behind it. Codex can also keep its own model and add Memtro as an MCP server. And a Codex credential under Providers lets Memtro answer codex/* models on your ChatGPT subscription, in a sandboxed container per request.

What you get

  • Custom model provider in ~/.codex/config.toml with base_url https://platform.memtro.com/v1 and wire_api = "responses"
  • Remote MCP server via codex mcp add, with default_tools_approval_mode = "approve" so Codex does not refuse every call
  • codex/* models in Memtro on a ChatGPT sign-in or an OpenAI API key, one sandboxed container per request

How to connect

  1. 01Create a personal API key on Dashboard → Profile → API keys and export it as MEMTRO_API_KEY
  2. 02Add a [model_providers.memtro] block to ~/.codex/config.toml, or run codex mcp add memtro --url https://platform.memtro.com/api/mcp --bearer-token-env-var MEMTRO_API_KEY

Example queries

  • › auto:agent as the model runs every turn as a multi-step investigation
  • › search memtro for what we decided about retries

Why it is better with memory

Through Codex 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.