Keva gives your AI assistants structured, layered, long-term memory of your project — decisions, constraints and knowledge — so they stop forgetting.
We are builders who got tired of re-explaining the same context to AI assistants every session. Keva is the memory layer we wished existed: structured, isolated per project, and fully under your control.
Keva stores your project knowledge in five cognitive layers and serves it back to any MCP-compatible AI client (VS Code Copilot, Claude Code, Gemini) exactly when it is needed.
Mission, project state, decisions, knowledge base and episodic history — weighted by intent.
Each project is fully isolated. Only you can see your data. No cross-tenant bleed.
A single connection token links VS Code, Claude Code or Gemini to your vault via MCP.
Runs on local embeddings — no API keys, no per-token billing for memory.
Create your free account in seconds.
Start free