Docmancer provides a persistent local memory layer for coding agents such as Claude Code, Cursor, and others. It uses RAG and local-first techniques so every agent on a machine shares one unified project memory, eliminating the need to repeatedly explain the same codebase details.
In the LLM eval & observability space, docmancer takes a focused approach. Repeatedly re-explaining project context to multiple local coding agents. It is built as an open-source project for developers. docmancer is open source under the MIT license. It runs on the web and the command line.
It is developed by Docmancer, and it first shipped in 2026. The project is developed in the open on GitHub with 116 stars and 130 commits in the last 90 days. Among its 4 catalogued features are Local Memory, Shared Context, and RAG. It exposes integrations via an MCP server.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do