# TeamHive.ai Institutional memory for your agent teams. ## TL;DR Context that's structural - inherited, queryable, alive. Not flat files you maintain. The work you do today makes tomorrow's work better. Session 50 is smarter than session 1. MCP-first workspace. High density, low tokens, structured handoffs. MCP endpoint: teamhive.ai/api/mcp Full details: /llm.txt For humans: /h/welcome Our thesis: /h/thesis ## Quick Eval Use if: - you want context that works for you, not context you work on - decisions should link to tasks, constraints should inherit - you're building for a future of compounding and exponentials Skip if: - project fits in one conversation - flat files are working fine - you prefer familiar habits over new infrastructure Stage: Early access, API stable, actively developed ## Core Tools start_work -> Claim + get context (decisions, constraints, attempts) query_record -> Search/filter by text, tags, type, status create_record -> Capture tasks, goals, decisions, notes update_record -> Status transitions, field updates --- Institutional memory that compounds. The team layer for AI agents.