Product positioning
Living Brain is an interoperable memory layer for agentic AI from ChatAndBuild. The official site describes it as a "persistent, self-learning memory layer" designed to keep knowledge connected across different AI tools and agents, so agents don't have to rebuild context from scratch every time.
Confirmed key features
- Persistent memory and knowledge compilation: Organizes scattered conversations, files, and updates into structured, source-linked knowledge.
- Multiple access methods: The official site states it can be used via REST API, SDK, CLI, and MCP, and supports connecting custom MCPs.
- Semantic search and knowledge discovery: Offers semantic search, connection discovery, Brain scoring and suggestions, daily briefings, and more.
- Context capture: Supports note and URL capture, PDF/DOCX/PPTX/HTML/text extraction, audio transcription, batch capture, and a Telegram capture bot.
- ChatChat workspace integration: Create, organize, and run team Brains and agents within ChatChat.
Use cases
Suited for developers and teams that need multiple AI tools or agents to share long-term knowledge—for example, giving agents retrievable project background, document knowledge, or historical decisions. It also fits workflows where you want models to adapt to your processes at inference time without modifying model weights. The official site also mentions compiling accumulated knowledge and feedback into training and evaluation datasets for post-training pipelines.
Pricing
The official pricing page shows:
- Living Brain API: Pay-as-you-go, starting with a $5 prepaid balance, no monthly minimum; model input $8/1M tokens, output $0.50/1M tokens.
- Living Brain Workspace: $99/month, including 3 team seats, 10 GB storage, 500,000 requests per month, and more.
- Business: Custom pricing based on deployment scope; the official site does not list a fixed price.
Specific quotas, rate limits, and feature inclusions are subject to the official pricing page.
Notes
Some benchmark data shown on the official site (such as LoCoMo and LongMemEval-S results) comes from its internal evaluations, and rankings apply only to the systems and configurations tested.