Find My AI Tool Logo
All tools

ContextNest

ContextNest is governed context for AI: a company's knowledge stored as structured, graph-linked Markdown that is versioned and approved by its owners, and read by any MCP-compatible AI such as Claude, Cursor, or Antigravity.

Developer ToolsBusiness ManagementEnterprise SoftwareAI AgentsAI GovernanceKnowledge Management

About ContextNest

ContextNest

ContextNest, from PromptOwl, gives every AI a team uses the same approved company knowledge and records every read, so teams can show what an AI was working from on a given date. Knowledge is stored as structured, graph-linked Markdown. Owners approve changes when review is turned on, every version is kept, and every read is recorded. Any MCP-compatible AI (Claude, Cursor, Antigravity) reads directly from the nest. The page describes starting with plain files via an open-source CLI, self-hosting the server in Docker, or using the hosted service. It positions ContextNest for handling living knowledge, in contrast to RAG for static, large corpora. The page also cites a research preprint (not peer reviewed) reporting a 97% answer pass rate versus 90–93% for BM25 keyword search in a stale-version test, at about a third of the input tokens. The Desktop app is stated to be free and run locally, and the product is described as patent pending.

Key features

  • Structured, graph-linked Markdown knowledge vault
  • Versioning of every document with version history and compare
  • Approval workflow where a named owner approves changes before AI may use them
  • Record of every read by AI and users
  • Wiki-style editor and built-in Hootie chat
  • Works with any MCP-compatible AI client such as Claude, Cursor, and Antigravity
  • Open-source CLI, engine, and MCP server (AGPL-3.0)
  • Deployment options: local plain files, self-hosted via Docker, or hosted cloud
  • OIDC single sign-on with Entra ID, Okta, or Google Workspace on Enterprise
  • Role-based access control
  • SOC 2 and HIPAA controls; SOC 2 reporting and HIPAA BAA on Enterprise

Best suited for

  • Teams that need consistent answers across multiple AI tools and agents
  • Organizations in regulated industries needing approval and audit records of AI knowledge
  • Agencies and consultancies deploying a method to client AI
  • Associations and member organizations offering assistants built on approved material
  • Enterprises governing AI knowledge across teams
  • Developers building agents that read a shared governed source

Pricing

The CLI, engine, and MCP server are open source and free. The team workspace is free for up to 10 users. Hosted is $10 per user per month. Self-Hosted is $10 per user with a 30-user minimum. Every paid plan starts with a free month. Enterprise is custom, on one annual contract.

AI agentsOpen sourceMCPAI governanceKnowledge managementAudit trailContext managementApproval workflowVersion controlSelf-hosted

Discussion

No comments yet. Start the conversation.