AGENT INFRASTRUCTURE · WORTH A LOOK

Living Brain

A memory layer from ChatAndBuild that ChatGPT, Claude, Cursor, Codex and your own agents can all write to and read from.

Visit livingbrain.com ↗@livingbrain on X ↗

What Living Brain does

Stores documents, conversations and decisions outside the model, compiles them into linked pages (a knowledge graph), and retrieves the relevant context for each task. It's available through a hosted MCP server, REST API, SDKs and a CLI, with ChatGPT and Claude connectors. The API is in early access.

Good for

Keeping one project memory across ChatGPT, Claude, Cursor and Codex
Giving your own agents persistent memory via MCP or an API
Cutting the tokens spent re-sending long histories

Try this in 10 minutes

Honest take

No red flags in our notes so far

A real, documented product aimed at a real gap: memory that isn't locked inside one assistant. The headline numbers (92.2% LoCoMo, 22.6x fewer tokens, 0% hallucinations) are all the company's own internal evaluations, and the launch post mixes figures from two different benchmarks. Worth a trial if you hop between assistants; wait for independent tests before trusting the numbers. We haven't tested it hands-on.

Our Hype or Fact? checks

Partly factDoes Living Brain really give every AI one shared memory, at 92.2% on LoCoMo with 22.6x fewer tokens? 4 Nov

Pricing

API: $8 per 1M input tokens and $0.50 per 1M output tokens, topped up from $5 with no subscription (1 GB storage, 100k requests a month). A team workspace comes with ChatChat Pro+, and Business is custom-priced.

Prices as listed by the maker when we last checked (2026-11-04); usually in US dollars. Check the website for current prices.

Latest update

4 November 2026: 2026-11-04: Added after the 4 Nov launch post from ChatAndBuild founder Christel Buchanan. API in early access, MCP, SDKs and a CLI, pay-as-you-go pricing published. Benchmarks are internal.