# Jess

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Sajiv Jess](https://cerebralvalley.ai/u/SajivJess)
- **GitHub:** https://github.com/SajivJess/MuseFlow---Meta
- **Demo video:** https://youtu.be/EUgokFpMYUQ
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3985

MuseFlow is a mobile-first AI engineering workspace that enables developers to control autonomous software development workflows directly from their smartphones.

Traditional AI coding tools focus primarily on generating code. MuseFlow explores a more complete approach: allowing developers to delegate real engineering tasks, define execution boundaries, monitor agent activity, verify code changes and retain final approval authority — all through a mobile interface.

The platform introduces a two-gate, human-governed workflow:

Gate 1 – Contract Before Code: Before implementation begins, the AI agent works with an execution contract containing task requirements, acceptance criteria, permitted files and verification commands. The developer reviews and approves this contract before execution.

Autonomous Execution: Meta-powered AI agents work on the selected project, modify code within the approved scope and report their progress.

Evidence-Based Verification: MuseFlow runs actual verification commands and captures execution results, exit codes and acceptance-criteria outcomes, distinguishing executed evidence from AI-generated claims.

Gate 2 – Human Approval: If verification fails, commit approval is blocked. When verification succeeds, the developer receives the authority to review the changes and approve the commit.

Built with Flutter, Python, FastAPI and Meta Muse Spark, MuseFlow brings mobile-first interaction, agentic software engineering and human accountability together.

Our vision is simple: developers should be able to control, verify and approve AI-driven software engineering from anywhere.

MuseFlow — YOUR IDE. EVERYWHERE.

## More from Meta Global AI Developer Hackathon

- [Kriah Listener](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3982)
- [Aridaq](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3983)
- [Meta Agent: Describe a problem, get a working AI agent (Muse Spark)](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3984)
- [Patient Operations Record](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3986)
- [Miracle Tsaka](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3987)
- [Keerthi Ananthula](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3988)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3985. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt