CogniMesh

CogniMesh is a proposed open-source project connected to Devman. It explores how domain experts can structure, host, and monetize their knowledge as self-owned RAG endpoints that AI agents can discover and query.

The project starts from a simple position: AI context should not depend only on centralized data monopolies. Universities, researchers, academics, and professional specialists should be able to run knowledge nodes without surrendering ownership of their source material.

Core Modules

CogniMesh is planned around three modules:

AI Interviewer Agent: a local or self-hosted agent that interviews domain experts and turns their answers into structured knowledge.
P2P Network and Service Discovery: a mesh where knowledge nodes can announce availability, categories, metadata, and quality signals.
Decentralized Monetization and Rating: payment and reputation mechanisms so providers can be compensated and consumer agents can choose trustworthy sources.

Start Here

Overview explains the problem and project motivation.
Architecture describes the initial system model.
Roadmap outlines the staged MVP direction.
NLnet Proposal Summary gives a sanitized public grant-context summary.