OtoDock appeared on Hacker News showing a self-hosted company OS that places Claude Code and Codex agents into separate departments.
The project reached 37 points and 9 comments in its first discussion thread.
What It Is and How It Works
OtoDock runs entirely on your own infrastructure. Each department contains one or more agents powered by Claude or Codex models. Departments handle distinct functions such as engineering, research, or operations and communicate through defined internal channels.
Agents receive tasks, maintain context within their department, and produce outputs that feed into other departments. The system uses standard container deployment so teams can spin up or shut down departments without external API calls.
Community Reception on Hacker News
The thread drew 37 points and 9 comments. Readers noted the appeal of keeping all agent activity inside company firewalls. Several asked about scaling limits when multiple departments run simultaneously.
Others questioned how the system resolves conflicting outputs between departments. The discussion stayed technical and focused on deployment details rather than hype.
How to Try It
Clone the repository and follow the provided Docker setup. The repo at https://github.com/OtoDock/oto-dock contains the core configuration files and example department templates.
Users assign API keys for Claude and Codex locally, then define department roles in YAML. No hosted service is required.
Pros and Cons
- Keeps all prompts and outputs inside your infrastructure
- Uses proven models (Claude, Codex) without building new ones
Simple department structure maps to existing company roles
Requires local GPU or API credits for each active agent
Limited public benchmarks on multi-department coordination
Still early-stage with only 37 HN points of feedback
Alternatives and Comparisons
| Feature | OtoDock | CrewAI | Auto-GPT |
|---|---|---|---|
| Self-hosted | Yes | Partial | Yes |
| Department structure | Built-in | Custom workflows | Single agent focus |
| Model support | Claude + Codex | Multiple | Multiple |
| HN discussion points | 37 | Higher volume | Higher volume |
OtoDock emphasizes fixed departments over dynamic task graphs found in CrewAI. It trades flexibility for clearer organizational mapping.
Who Should Use This
Teams already running self-hosted LLMs and wanting explicit departmental separation will find the structure useful. Companies with strict data residency rules gain the most.
Skip it if you need rapid prototyping or already rely on managed agent platforms with built-in orchestration.
Bottom Line / Verdict
OtoDock offers the first explicit department-based layout for Claude and Codex agents that stays fully self-hosted. Early adopters gain organizational clarity at the cost of managing their own infrastructure.
The approach suits mid-size teams that treat AI agents as permanent staff rather than temporary scripts.
Top comments (0)