Agents: for developers and agent builders
One command. Plain files you can read.
Start governed context on your machine with one command, connect it to Claude Code or Cursor, and take it to your team when it works. The command-line tool, engine and MCP server are open source.
$ npx @promptowl/contextnest-cli init
The moment
You found the tool in a thread, on GitHub or on npm, and you want to know if it beats what you've already wired up.
What you've probably triedA memory library, a folder of markdown, a vector store and some glue code.
What people in your role ask
- “Is this just another memory tool?”
- “Can my agents in Claude Code and Cursor share it?”
- “Is it plain files, or a black box?”
- “How do I get from works-for-me to my team can use it?”
Where it sits
Where governed context sits in your stack.
Layer 3
Agents, models and runtimesSwappable. A run can end when its job is done.
Claude
Cursor
PromptOwl agents
Gemini
Read, scoped by rolePropose updates, an owner approves
Layer 2
Governed context layerContextNest. Stays put.
The company's knowledge, in plain files it owns, that every agent reads.
ApprovedCurrentScoped to each jobVersionedEvery read recorded
Distill and approve what agents may use
Layer 1
Where information livesStays where it is. Agents don't read it directly.
Google DriveSharePointEmailCRMMeeting transcripts
What you'll want to see
The proof that matters to you.
Plain text files
Markdown you can read, diff and keep without us.
A fair comparison with memory tools
Where Mem0, Zep and built-in AI memory win, and where they stop.
Read more →Open source where it counts
The command-line tool, engine and MCP server are open source under AGPL-3.0. The desktop app and team server are separately licensed.
Read more →A path to your team
Hosted or self-hosted, with approvals and scoping when you need them.
Read more →Start where you are.
Start free with the AI tools you already use, or book a demo on documents like yours.