--- title: I Built an Open-Source Agent OS published: true description: "A locally-hosted multi-agent OS that orchestrates opencode, Hermes Agent, and Gemini CLI into one dashboard with 16 skills, scheduler, cost analytics, and memory. Full architecture breakdown." tags: [agents, devops, opensource, python, fastapi] cover_image: https://raw.githubusercontent.com/modimihir07/agentic-os/main/docs/og-image.png canonical_url: https://dev.to/mihir_nmodi_14a06a4019e1/i-built-an-open-source-agent-os-2h30 --- I built an operating system for AI agents. Not a terminal. Not a chat app. A full dashboard that coordinates three agents — **opencode** (code/DevOps), **Hermes Agent** (memory/scheduling), and **Gemini CLI** (research) — into one unified control plane. GitHub: [github.com/modimihir07/agentic-os](https://github.com/modimihir07/agentic-os) --- ## The Problem I was jumping between three tools: - opencode for coding tasks - Hermes for scheduling and memory - Gemini CLI for web research No shared context. No central dashboard. Every session started from zero. ## The Solution: 7-Layer Architecture ``` Layer 7 Identity / Constitution Layer 6 Self-Evolution Layer 5 Scheduler + Health Layer 4 Memory Graph Layer 3 Skills Hub + Eval Layer 2 Business Brain Layer 1 Agent Router ``` **Layer 1** routes tasks automatically: code → opencode, memory → Hermes, research → Gemini. **Layer 3** gives each skill a standardized structure (SKILL.md + learnings + eval + score history) so they improve over time. **Layer 4** provides persistent memory via a shared `brain/` folder and SQLite FTS5 — agents remember context across sessions. ## 12 Features | Feature | What it does | |---------|-------------| | 3-Agent Engine | opencode + Hermes + Gemini CLI with auto-routing | | 16 Skills | Executable packs with eval scoring | | Cron Scheduler | APScheduler jobs (heartbeat, standup, audit) | | Cost Analytics | Track spending across providers | | One-Click Backup | Full tar.gz snapshots | | Audit Trail | Every action logged | | Prompt Library | 10 reusable templates | | Dark/Light Theme | GitHub-style dark mode | | Standards System | Auto-discover project conventions | | Plugin Registry | Extend via custom skills | | Client Timeout | 200s AbortController for long queries | | Memory | SQLite FTS5 + shared brain/ files | ## vs Claude Agent OS Honest comparison with Julian Goldie's video project: | Feature | Claude Agent OS | Agentic OS (Mine) | |---------|----------------|-------------------| | Agents | Claude + OpenClaw + Hermes | opencode + Hermes + Gemini | | Cost | $20/mo | **$0 (all free tiers)** | | Memory | Obsidian (external) | **Built-in** | | Cost Tracking | None | **Built-in** | | Backup | None | **One-click** | | License | Tutorial only | **MIT** | ## Tech Stack Backend: FastAPI (Python) · Frontend: vanilla JS SPA · Scheduler: APScheduler · Memory: SQLite FTS5 · Cost: $0 (Gemini Flash, OpenRouter free models) ## Quick Start ```bash git clone https://github.com/modimihir07/agentic-os.git cd agentic-os && ./install.sh && ./start.sh # Open http://127.0.0.1:8080 ``` --- Agentic OS is MIT licensed. Star it on GitHub if you find it useful. 👉 **[github.com/modimihir07/agentic-os](https://github.com/modimihir07/agentic-os)**