Atris
Run Atris campaigns from any agent, and extract, check, or search brand design systems.
At a glance
Not tested yet- GitHub stars
- 71
- Weekly downloads
- 7,514
- Last commit
- 2026-11-19
- Licence
- MIT
- Transport
- stdio
- Auth
- api-key
npx -y atris
What it does
Atris is an operating system for AI work that turns a repo into a workspace an agent can operate. It gives coding agents a memory that lasts across sessions, stored as plain files in your repo. It runs a plan -> do -> review loop with durable tasks, daily logs, and verification that ends work on a real check. It is for people and teams who want to see what their AI actually did.
Key features
- A shared memory in plain files:
atris/MAP.md,atris/wiki/,atris/logs/, andatris task, read by every agent before it acts. - A strict
plan -> do -> reviewloop with daily logs, task tracking, feature packs, and project memory. - Persistent AI computers for scoped jobs, personal or shared by a business.
- Members that work in their own copy of the repo, with run, wake, sleep, and result commands.
atris recapgives a plain answer to what the AI did, andatris recap --sharegives a receipt-backed summary.atris aeo scanchecks whether answer engines can read your site, with no login needed.
Quick start
npm install -g atris cd your-project atris init && atris
Then read the workspace's atris/atris.md and follow it exactly.
Add this to .mcp.json in your project root (Claude Code) or .cursor/mcp.json (Cursor):
{
"mcpServers": {
"atris": {
"command": "npx",
"args": [
"-y",
"atris"
],
"env": {
"ATRIS_API_KEY": "YOUR_ATRIS_API_KEY"
}
}
}
}Replace each YOUR_ value with your own, and delete the variables you do not use.
Environment variables
ATRIS_API_KEY: not explained in the sources.
Example prompts
- Fix one failing test and show me the proof.
- What did my AI actually do today?
- Check whether ChatGPT can read my site and compare it to a competitor.
What to watch out for
- Requires Node.js 18+.
atris taskuses built-in SQLite and requires Node.js 22+. - Atris cloud workspaces, businesses, and integrations need
atris setupafter install. - Work lands on its own only after you switch self-landing on once with
atris autoland on. - A job that fails three times in a row stops and waits for a person.
Not tested by tools4agents yet. Read the source repository before you install it.
Similar tools
Blender MCP Server
Connects Blender to AI coding assistants via MCP for prompt-assisted 3D modeling, material creation, and scene inspection.
Pascal 3D Editor
Create, inspect, validate, and save editable 3D building scenes with Pascal's MCP server.
Figma Context MCP Server
Gives coding agents direct access to Figma component structures, tokens, and layouts for accurate UI implementation.
Unity MCP Server
Connects Unity Editor to AI coding agents to control scene hierarchies, edit scripts, inspect game objects, and manage assets.