DocuMemory: A Living Memory Engine for AI-First Engineers

Summary

DocuMemory is a developer tool that turns AI chat sessions into persistent architecture decisions, engineering documentation, and reusable context blocks for later sessions. It is designed to work with Claude Code, ChatGPT, Cursor, Windsurf, and other MCP-compatible clients. The product captures a conversation through an MCP function, then applies subtractive distillation to extract facts, remove noise, and validate schemas. Its example output is a structured memory document containing project metadata, key technical facts, and a compact AI context snippet. Recall combines semantic and full-text search through reciprocal rank fusion, then injects the selected context back into the AI client. The service emphasizes MCP-native access, sub-50-millisecond hybrid search, and local-first or self-hosted deployment using the customer’s infrastructure, Postgres, and AI keys; it also describes using a local model to keep knowledge on the user’s machine. Pricing includes 15 free distillations, a $12 monthly individual plan, a $10-per-month equivalent annual plan, and a team plan priced at $15 per seat monthly with a two-seat minimum and no free trial.