Odysseus is an open-source, self-hosted AI workspace that provides ChatGPT/Claude-like functionality on your own hardware. It includes chat, AI agents with tool use, deep research, code editing, email/calendar integration, and more—all running locally with full data privacy.

How is Odysseus different from ChatGPT or Claude?

Unlike cloud-based services, Odysseus runs entirely on your infrastructure. You control your data, choose your models (local or API), avoid subscription fees, and get features like agentic tool use, deep research, email management, and calendar integration that aren't available in consumer AI products.

Can I run Odysseus on my own computer?

Yes. Odysseus supports Docker (recommended), native Linux/macOS, Windows, and Apple Silicon with GPU acceleration. It works with local models via Ollama/llama.cpp/vLLM or cloud APIs from OpenAI, Anthropic, and others. The Cookbook feature even recommends models based on your hardware.

Is Odysseus really free and open source?

Yes, Odysseus is MIT-licensed open source. You can use, modify, and distribute it freely. The only costs are your infrastructure (your computer or cloud server) and any API usage if you choose to connect cloud models instead of running local ones.

What AI models does Odysseus support?

Odysseus supports any OpenAI-compatible API, including: Ollama, llama.cpp, vLLM, OpenRouter, OpenAI, Anthropic Claude, and more. The Cookbook feature scans your hardware and recommends models that will run efficiently on your GPU/CPU.

Odysseus: The Self-Hosted AI Workspace That's Taking | explainx.ai Blog

explainx.ainewsletter3.5k

workshops ↗

Odysseus: The Self-Hosted AI Workspace That's Taking | explainx.ai Blog | explainx.ai

Feature	ChatGPT	Claude	Odysseus
AI agents with shell access	❌	Limited	✅
Email integration	❌	❌	✅ (IMAP/SMTP/CalDAV)
Calendar management	❌	❌	✅
Deep research mode	❌	❌	✅
Local model serving	❌	❌	✅ (Cookbook)
Persistent memory/skills	Limited	Limited	✅ (ChromaDB)
Multi-tab document editor	❌	❌	✅
Model comparison (blind)	❌	❌	✅

Tool Category	Capabilities
MCP (Model Context Protocol)	Connect to any MCP server (file systems, databases, APIs)
Web search	Query SearXNG (bundled), retrieve and parse results
File operations	Read, write, edit files on your system
Shell access	Execute bash commands (with safety controls)
Skills	Custom agent behaviors and workflows
Memory	Retrieve and store long-term context via ChromaDB

bash

Detected: NVIDIA RTX 4090 (24GB VRAM)
Recommended models:
- Llama 4 70B FP8 (fits with 80% utilization)
- Qwen 3 72B AWQ (fits with 70% utilization)
- DeepSeek V3 GGUF Q4_K_M (fits with 90% utilization)

snippet

User interaction → Extract important facts
              ↓
         Store in ChromaDB
              ↓
    Retrieve when relevant in future
              ↓
         Agent uses context

Feature	Description
IMAP/SMTP	Connect any email account
AI triage	Auto-categorize by urgency
Summaries	One-line summaries of each email
Auto-tag	Apply labels based on content
Auto-reply drafts	Generate response drafts for you to edit
Spam detection	AI-powered spam filtering
CalDAV-aware	Understands calendar invites

bash

git clone https://github.com/pewdiepie-archdaemon/odysseus.git
cd odysseus
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python setup.py
python -m uvicorn app:app --host 127.0.0.1 --port 7000

powershell

git clone https://github.com/pewdiepie-archdaemon/odysseus.git
cd odysseus
py -3.11 -m venv venv
venv\Scripts\Activate.ps1
pip install -r requirements.txt
python setup.py
python -m uvicorn app:app --host 127.0.0.1 --port 7000

Setting	What to Add
Models	Add local models (Ollama, vLLM, llama.cpp) or cloud APIs
Search	Configure SearXNG or use bundled Docker instance
Email	Add IMAP/SMTP credentials
Calendar	Add CalDAV URL
MCP Servers	Add custom MCP integrations
API Tokens	For webhook/API access

bash

# App binding (127.0.0.1 for local, 0.0.0.0 for network)
APP_BIND=127.0.0.1
APP_PORT=7000

# Auth (keep true for network deployments)
AUTH_ENABLED=true
LOCALHOST_BYPASS=false  # Dev only
SECURE_COOKIES=false    # Set true when behind HTTPS proxy

# Database
DATABASE_URL=sqlite:///./data/app.db

# ChromaDB (memory)
CHROMADB_HOST=localhost
CHROMADB_PORT=8100

# SearXNG (search)
SEARXNG_INSTANCE=http://localhost:8080

# Model hosts (optional, or add in UI)
LLM_HOST=localhost
OPENAI_API_KEY=sk-...

Layer	Technology
Backend	FastAPI (Python)
Frontend	Vanilla JS + HTML/CSS (modular)
Database	SQLite (sessions, messages, documents)
Vector DB	ChromaDB (memory)
Search	SearXNG (bundled in Docker)
Notifications	ntfy (optional, bundled)
Model serving	vLLM, llama.cpp, Ollama

snippet

app.py                   # FastAPI entry point
core/                    # auth, database, middleware
src/                     # llm_core, agent_loop, agent_tools, chat_processor
routes/                  # chat, session, document, memory, model endpoints
services/                # docs, memory, search, hwfit (Cookbook)
static/                  # index.html + app.js + style.css + js/
docs/                    # landing page + preview clips
data/                    # user data (app.db, memory, uploads, etc.)

Feature	Odysseus	Open WebUI	Jan.ai	LibreChat
Agent tools	✅ Full (MCP, shell, web)	⚠️ Limited	❌	⚠️ Limited
Deep research	✅	❌	❌	❌
Email integration	✅	❌	❌	❌
Calendar	✅	❌	❌	❌
Document editor	✅	⚠️ Basic	❌	⚠️ Basic
Memory (vector)	✅ ChromaDB	⚠️ Basic	⚠️ Basic	❌
Cookbook	✅	❌	✅	❌
Multi-model	✅	✅	✅	✅

Limitation	Impact	Workaround
No mobile app	Web-only (though PWA installable)	Use browser, works well on mobile
Single-user focused	Multi-user works but limited	Deploy per-user or use admin controls
Local models need GPU	CPU inference slow for large models	Use smaller models or cloud APIs
Initial setup complexity	Requires technical knowledge	Use Docker for simplicity

Clone the repo:

bash

git clone https://github.com/pewdiepie-archdaemon/odysseus.git
cd odysseus

Choose your method:
- Docker: docker compose up -d --build
- Native: ./start-macos.sh or follow Linux/Windows instructions
- Apple Silicon: ./start-macos.sh for GPU support
Open in browser:
- http://localhost:7000 (Docker/Native)
- http://127.0.0.1:7860 (macOS script)
Log in with admin credentials (printed in terminal)
Configure in Settings:
- Add models (Ollama, OpenAI, etc.)
- Set up search, email, calendar
- Install MCP servers
Start chatting, coding, researching!

GitHub Stars	22.4K
Forks	2.8K
Contributors	88
License	MIT (fully open source)
Primary Languages	JavaScript (47.7%), Python (38.9%), CSS (10.4%)
First Release	v1.0 (June 2026)
Deployment Options	Docker, Linux, macOS, Windows, Apple Silicon

What Is Odysseus?

The Elevator Pitch

Key Stats

Related posts

Meetily: Privacy-First AI Meeting Assistant With Local Whisper and Parakeet

video-use: Edit Videos With Claude Code — No Premiere Pro Needed

Google Fired the Engineer Who Built Its Viral Workspace CLI — Two Days Before Announcing the Official One

Why Odysseus Matters

1. Data Privacy and Control

2. No Subscription Fees

3. Model Flexibility

4. Features Not Available Elsewhere

Core Features Deep Dive

1. Chat: Multi-Model Conversations

2. Agents: Tool-Using AI Assistants

3. Deep Research: Multi-Step Information Synthesis

4. Cookbook: Model Management Made Easy

5. Documents: AI-Assisted Editing

6. Memory & Skills: Persistent Context

7. Email: AI-Powered Inbox Management

8. Calendar: Local-First Scheduling

9. Notes & Tasks: Quick Capture

Installation and Setup

Docker (Recommended)

Native Linux/macOS

Apple Silicon (GPU Acceleration)

Windows

Configuration

Essential Settings

Environment Variables (.env)

Security Best Practices

Architecture

Tech Stack

Project Structure

Data Storage

Use Cases

1. Privacy-Conscious Development

2. Enterprise On-Premises AI

3. Research and Analysis

4. Personal Knowledge Management

5. Multi-Model Experimentation

Community and Ecosystem

GitHub Presence

Contributing

Alternatives Comparison

Limitations and Considerations

Current Limitations

Hardware Requirements

The Future of Odysseus

Roadmap Highlights

Bottom Line

Getting Started

Related Posts