What problem it solves

Teams that rely on AI agents need a way to give each agent a role, workspace, and instructions, keep context across conversations, and control what agents are allowed to do. Without that, agents run blind: actions happen without review, memory does not persist, and repeating scheduled work means redoing setup by hand.

Key features

• Create agents and teams — Give an agent a name, role, engine, workspace, memory, and approval settings, and let a lead agent delegate work to specialists or coordinate agents in shared rooms.

  Source: README, docs/TEAMS.md

• Keep useful context with memory — Agents save notes, recall relevant notes and conversations, and search their history, while you can review and edit what they remember.

  Source: README, docs/MEMORY.md

• Run scheduled routines — Create routines for scheduled, webhook, or manual runs and review each result; a routine never approves an action by itself.

  Source: README, docs/ROUTINES.md

• Approval modes and visibility — Pick Ask first, Accept edits, Approve for me, or Plan only, then review tool activity, approval decisions, and pending items in the Today view.

  Source: README, docs/EVERYDAY.md

• Bring your own engine and apps — Connect Claude Code or Codex CLI, Gemini CLI, Cursor, ACP agents, OpenRouter, or an OpenAI-compatible API, and connect services such as GitHub, Gmail, and Google Calendar through your own Composio key.

  Source: README, docs/ENGINES.md

Quick start

Copied verbatim from the project's README.

git clone https://github.com/Anil-matcha/open-duet.git
cd open-duet/desktop
npm run app

Who it may suit

This may suit individuals and small teams who want to run named AI agents on their own machine or server with explicit approval gates, especially those willing to supply their own engine and provider account.

Topics to discuss

• Which approval modes fit real workflows, and where does the friction of pausing for review become a problem?

• What should the shared conventions be for agent memory notes, skills, and templates so work stays consistent across a team?

• How should self-hosted deployments handle scheduling, backups, and upgrades once routines and agent computers are used for recurring work?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/Anil-matcha/open-duet#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.