Odysseus: Self-Hosted AI Workspace with 10+ Built-in Tools

Odysseus (69,110 GitHub stars) is a self-hosted AI workspace combining chat, agent automation, deep research, document editing, email triage, calendar, and more. Supports vLLM, llama.cpp, Ollama, OpenRouter, OpenAI, and GitHub Copilot. Docker and native Linux/macOS installs available.

  • ⭐ 81348
  • Updated 2026-06-09

Odysseus: The Self-Hosted AI Workspace That Hit 63open-notebook: The Open-Source Notebook LM Alternative That Supports 15+ AI Providers — Self-Hosted, 28,000 Stars — Setup Guide 2026

┌──────────────────────────────────────────────────┐
│              Odysseus Architecture                 │
│                                                   │
│  ┌─────────┐  ┌──────────┐  ┌─────────────────┐  │
│  │  Chat   │  │  Agent   │  │    Cookbook     │  │
│  │ (API)   │  │ (Tools)  │  │ (Model Server)│  │
│  └────┬────┘  └────┬─────┘  └────────┬────────┘  │
│       │             │                  │           │
│       ▼             ▼                  ▼           │
│  ┌───────────────────────────────────────────────┐ │
│  │           Python Backend (FastAPI)             │ │
│  │  ChromaDB │ SearXNG │ ntfy │ .env config    │ │
│  └───────────────────────────────────────────────┘ │
│       │             │                  │           │
│       ▼             ▼                  ▼           │
│  ┌───────────────────────────────────────────────┐ │
│         Docker Compose Stack                     │ │
│  ┌────────┐ ┌──────────┐ ┌────────┐ ┌─────────┐ │
│  │ Odysseus│ │ChromaDB  │ │SearXNG │ │ ntfy    │ │
│  └────────┘ └──────────┘ └────────┘ └─────────┘ │
│                                                   │
│  ┌───────────────────────────────────────────────┐ │
│  │        Frontend: Responsive Web UI (PWA)      │ │
│  └───────────────────────────────────────────────┘ │
└──────────────────────────────────────────────────┘

Odysseus is a self-hosted AI workspace that brings together 10+ integrated tools into a single, privacy-first interface. Created by the developer known as pewdiepie-archdaemon, it has already surpassed 65,000 GitHub stars since its creation on May 31, 2026 — one of the fastest-growing AI projects in GitHub history.

Unlike ChatGPT or Claude which require you to hand over your data, Odysseus runs entirely on your own hardware. You connect your own API keys or serve local models yourself. The project describes itself as “the self-hosted version of the UI experience you get from ChatGPT and Claude, but with more jank and fun.”

This guide covers everything: architecture breakdown, Docker and native installation, model configuration, agent setup, deep research, and production hardening.

Get a DigitalOcean account for running this at scale

What Is Odysseus? #

Odysseus is a full-stack AI workspace built on Python (FastAPI backend, responsive web frontend). It is designed for users who want the convenience of a unified AI interface — like ChatGPT’s multi-model chat — while maintaining complete data sovereignty.

The project integrates the following capabilities in a single web application:

FeatureDescriptionBuilt On
Multi-model chatChatGPT-style interface for any connected modelFastAPI + WebSocket
Agent automationTask execution with tool useAgent loop + tools
Deep researchMulti-step research with citationsSearXNG + ChromaDB
Document editingIn-app document creation and editingWeb editor
Email triageInbox summarization and draftingIMAP integration
CalendarSchedule and event managementCalendar integration
Vector memoryPersistent context across sessionsChromaDB

Installation #

git clone https://github.com/pewdiepie-archdaemon/odysseus
cd odysseus
cp .env.example .env   # configure your API keys / model endpoints
docker compose up -d

This starts the full stack: Odysseus web app, ChromaDB (vector memory), SearXNG (search), and ntfy (notifications).

Native Linux/macOS #

git clone https://github.com/pewdiepie-archdaemon/odysseus
cd odysseus
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
uvicorn main:app --host 0.0.0.0 --port 8080

Model Configuration #

Odysseus supports multiple backends — configure in .env:

# Local via Ollama
OLLAMA_BASE_URL=http://localhost:11434
DEFAULT_MODEL=llama3.1:8b

# Or hosted via OpenRouter
OPENROUTER_API_KEY=sk-...
DEFAULT_MODEL=anthropic/claude-sonnet-4

# Or OpenAI-compatible
OPENAI_API_KEY=sk-...
DEFAULT_MODEL=gpt-4o

You can switch models per conversation in the UI.

Agent & Deep Research #

Odysseus agents can chain tools:

  1. Search the web via SearXNG
  2. Read and summarize pages
  3. Store findings in ChromaDB vector memory
  4. Compose a cited report

This turns a single prompt like “research the current state of open-source RAG frameworks” into a multi-step research pipeline with sources — all running locally.

Production Hardening #

  • Put it behind Caddy or Nginx with automatic HTTPS
  • Set SECRET_KEY and disable debug mode
  • Back up .env and the ChromaDB data directory
  • Use ntfy to get push notifications for long-running agent tasks
  • Pin model versions for reproducibility

Conclusion #

Odysseus is the fastest-growing self-hosted AI workspace of 2026 for good reason: it combines the polish of a commercial AI product with complete data sovereignty. For privacy-conscious teams and individuals who want chat, agents, research, and documents in one place — without sending data to third parties — Odysseus is a compelling one-command deploy.

💬 Discussion