# Hermes Helmet by Machine Wisdom AI

> Hermes Helmet is Machine Wisdom AI's open-source software factory around Nous Research's Hermes Agent. It carries an accepted GitHub plan through implementation, review, repair, and authorized merge, with controlled autonomy and Hermes working under its own identity.

Machine Wisdom AI is led by Younes Abouelnagah, a Toronto-based applied ML systems engineer and fractional AI lead. The company site focuses on production AI, agent infrastructure, safety, evaluation, governed autonomy, and agent memory. Hermes Helmet is that company's product. It is not Nous Research, and it is not FAVA Trails.

The product tagline is: Give Hermes the wings to carry your plan through, under its own identity.

Hermes Helmet is for people already using coding agents who want to delegate a written plan without coordinating every task, review, and repair.

You put together a written plan with your coding agent and record the work as linked GitHub issues. The human Captain agrees on that plan and delegates authority. The coding-agent first officer uses the Captain's credentials to assign work, review it, direct repairs, and merge when authorized. Hermes implements, tests, opens pull requests, and repairs the same change under its own identity.

Codex or Claude Code can stay on planning and review while Hermes uses a separately configured supported model, including hosted open-weight options. That split is a configuration choice. It does not promise quality or measured savings. Changing the worker model does not change its identity or give it merge authority.

Optional FAVA Trails and OpenViking keep shared project knowledge available across connected clients. Delivery works without them.

The Apache-2.0 preview is documented in the published runtime image guide. Use that maintained page rather than pinning a particular image digest or model release here.

## Product and company

- [Hermes Helmet homepage](https://hermes-helmet.ai/): Public product site.
- [Machine Wisdom AI](https://machine-wisdom.ai/): Company site.
- [Machine Wisdom writing index](https://machine-wisdom.ai/writing/): Published writing.
- [Machine Wisdom AI llms.txt](https://machine-wisdom.ai/llms.txt): Company overview for readers and AI assistants.

## Product documentation

- [Hermes Helmet repository](https://github.com/MachineWisdomAI/hermes-helmet): Public source and issues.
- [README](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/README.md): Product overview and documentation map.
- [Quickstart](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/docs/quickstart.md): Install the host CLI and start the published worker image.
- [Captain, first officer, and crew](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/docs/captain-and-crew.md): Roles, delivery contract, and merge authority.
- [Model lanes](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/docs/model-lanes.md): Supported providers and worker model configuration.
- [Published runtime image](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/docs/runtime-image.md): Apache-2.0 preview documentation.
- [FAVA Trails integration](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/docs/fava-trails.md): Optional governed memory.
- [OpenViking integration](https://raw.githubusercontent.com/MachineWisdomAI/hermes-helmet/main/docs/openviking.md): Optional shared working context.

## Published writing and delivered proof

- [Hermes Helmet: giving your software factory wings](https://machine-wisdom.ai/writing/hermes-helmet-software-factory/): Published Machine Wisdom launch article.
- [FAVA Trails requirements issue 107](https://github.com/MachineWisdomAI/fava-trails/issues/107): Written plan for the installation and local-memory work.
- [FAVA Trails pull request 111](https://github.com/MachineWisdomAI/fava-trails/pull/111): Implementation, review, and same-PR repair.
- [FAVA Trails v0.7.0](https://github.com/MachineWisdomAI/fava-trails/releases/tag/v0.7.0): Published result of that work.

Try Hermes Helmet through the [quickstart](https://github.com/MachineWisdomAI/hermes-helmet/blob/main/docs/quickstart.md): send one GitHub issue through it with your preferred first officer and worker model.