Living Brain

Living Brain stores information outside an AI model and makes relevant pieces available during later tasks. Its publisher describes a shared memory for teams and agents. The product compiles conversations, files and updates into source-linked knowledge that a person or connected tool can retrieve.

What a shared brain changes

A team can link information and use the relevant brain from supported tools or agents. The service offers API, SDK and MCP access. The intended benefit is to avoid repeatedly supplying an entire conversation or document set with each request. A source link can also help a user see where a retrieved item came from. This is useful when different people or tools need the same background information across several sessions.

How the service measures its results

Living Brain publishes internal results for LoCoMo and LongMemEval-S. The company says its LoCoMo run achieved 464 milliseconds of median retrieval and 3.7 seconds of median total answer time. Those are different measurements. The latter includes generation, while retrieval covers search and fetching pages. The vendor separates the benchmark configurations and baselines, so a team should test its own documents and typical requests before using those numbers to forecast response time or cost.

Pricing and first setup

The publisher’s API pricing starts from a prepaid balance with a $5 minimum top-up and no monthly minimum. It lists $8 per million model input tokens and $0.50 per million output tokens in USD. Start with one shared subject and record what each tool should be able to retrieve. Then check the returned sources against the underlying documents, including what happens when an expected detail is missing.