# Kleos9

- **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:** [Manu R](https://cerebralvalley.ai/u/manu_r)
- **GitHub:** https://github.com/manuu-r/kleos9/
- **Demo video:** https://youtu.be/RbVmgWkvqpU
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3426

Building a hardware product requires multiple disciplines working together: mechanical CAD, PCB design, firmware development, and a companion app. Each typically requires its own tools and specialists. The biggest challenges often arise at the interfaces between them, such as pin mappings, enclosure fit, component compatibility, and BLE communication contracts.

Generic AI assistants don't fully solve this problem. They can generate plausible but unverified outputs, with no reliable mechanism to prevent an incorrect PCB design, incompatible component, or poorly fitting enclosure from being treated as a finished product.

Kleos9 is an agentic hardware and software development workbench. You describe the device you want to build, refine an interactive concept, evaluate its feasibility, and approve the shared engineering interfaces before development begins.

From there, specialized, isolated agents handle different aspects of development:

- Mechanical design: Generate actual STEP/STL files using build123d, along with rendered previews.
- Electronics: Create KiCad schematics and PCB layouts, route boards using Freerouting, run ERC/DRC checks, and export manufacturing files.
- Firmware: Generate and compile firmware for qualified hardware boards.
- Companion app: Develop a Flutter application with a web build.

However, generating these deliverables doesn't mean the product is ready for manufacturing or shipping.

Physical fabrication, component purchasing, 3D printing, assembly, packaging, and shipping are currently simulated. Firmware behavior on actual hardware and BLE communication with the companion app remain unverified. A successful concept preview or software validation is not equivalent to manufacturing approval.

This is where external agents and real-world execution services become essential. We need to integrate services such as Muse for tracking and coordinating ongoing tasks, and RentAHuman for hiring people to perform physical work that AI agents cannot execute themselves.

For example, an agent could coordinate PCB fabrication, hire a technician to inspect the manufactured board for defects, arrange physical assembly, and request real-device testing. The generated firmware must be flashed and validated on actual hardware, while the companion app must be tested for reliable BLE communication before the product can be approved for shipping.

The goal is to move beyond AI-generated designs toward an end-to-end, agent-orchestrated hardware development workflow, where AI handles design, engineering, coordination, and verification wherever possible, while humans and external services complete the necessary physical tasks.

Ultimately, Kleos9 aims to bridge the gap between describing a hardware product and delivering a physically tested, functional device, without confusing generated artifacts or simulated processes with real-world verification.

## More from Meta Global AI Developer Hackathon

- [Aeglos](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3423)
- [LosCommosHermanos](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3424)
- [LinguaCheck](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3425)
- [Minh Trinh](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3427)
- [OwlOwlOwl](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3428)
- [LUMI](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3429)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3426. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt