Your agents. A clear path forward.

Code Flow brings coding CLIs into one workspace for planning, implementation, review and verification. Start with a small task, follow the evidence, and decide when the result is ready to ship.

Try the preview

Start Code Flow locally and open it in your browser:

npx --yes --package=https://github.com/midhunxavier/code-flow/releases/download/v0.1.0-preview.7/midhunxavier-code-flow-0.1.0-preview.7.tgz codeflow

Use Node.js 22.13+, Git 2.38+ and a supported coding CLI installed and signed in. Runs on macOS, Linux and Windows through WSL2. Keep the terminal open while you work; Ctrl+C stops the app.

Prefer a macOS app? Download 0.1.0-preview.7:

• Download for Apple Silicon (M1 or later)
• Download for Intel Mac

Requires macOS 13+. The app opens its own window and keeps agents running when you close it. Node is included; Git, a signed-in coding CLI and Chrome or Edge for the shared browser are still required. The DMGs are unnotarized preview builds and macOS may block them; use npm if blocked.

Installation and setup · Release notes and checksums

From task to reviewed result

1. Open a project. Choose Open repository, confirm its project, then describe the outcome and add context.
2. Choose the team. Assign roles yourself or review a proposed Auto-team plan.
3. Follow the work. Answer questions, inspect changes and watch the checks.
4. Review and ship. Read the evidence before merging or opening a pull request.

Captured from the local application with demo agents. See screenshot provenance.

Work with agents

• Team patterns · Auto team · Project missions
• Runs and the workspace · Review inbox · Verification
• GitHub and shipping · GitHub review threads
• Selective task reruns · Deferred starts

Make the workspace yours

• Configuration · ACP agents · Context inspection
• Project conventions · Checkpoints and reversions
• Service launch profiles · Code intelligence · Architecture explorer
• Project memory and knowledge graph · Semantic claim review
• Coordinated repository changes · Model training preview

Connect and collaborate

• Project integrations
• Team collaboration · Shared project accounts
• SSH projects · Remote workers
• GitLab and Bitbucket workflows · External sources
• Automations and remote access · Notifications
• VS Code integration

Operate safely

• Security and execution boundaries · Filesystem policies
• Agent containers · Filtered Docker tasks
• Dependency discovery · Backup and maintenance

Build and extend

• Architecture and development · API and CLI · Extensions
• Provider compatibility · Accessibility
• Performance · Run storage
• Documentation and GitHub Pages

Engineering records

These documents record implementation decisions and evidence at a particular point in time. Use the guides above for current product instructions.

• Application improvement program
• Project navigation review and implementation plan (proposed)
• Dynamic-team evaluation and implementation plan
• Implementation review and original plan review
• Standing-team implementation plan
• Local and remote ML automation implementation plan
• Measured performance

Project README · Getting started

Project collaboration lives in More → Board and More → People. Scheduled tasks, backlogs and missions live in More → Automation; configured Board and Automation tabs stay visible. Global Inbox and Scheduling cover the current computer.