Everything AI Should Be.
Free.
A desktop application that gives you AI chat, a built-in browser, specialist agents, visual workflows, and persistent memory. No subscriptions. No lock-in. Open source.
AI tools cost too much for what they do.
You pay monthly for chat interfaces, limited agents, and cloud-locked data. Stop paying rent on software that should be yours.
Chat + limited tools. Cloud only.
Chat + projects. No local option.
Code editor + AI. Developer only.
All of the above. Combined. Free. Yours.
Built different. On purpose.
No Subscription
No monthly fees. No usage limits. No premium tiers. Download it, use it, keep it. You only pay your LLM provider directly — at their cost, not ours.
Your Machine, Your Data
Everything runs natively on your computer. Conversations, files, agent memory — nothing leaves your machine. Built-in P2P encrypted chat means even your messages never touch a server.
No Walled Garden
Full source code. AGPL-3.0 license. Fork it, modify it, extend it, redistribute it. We don't lock features behind paywalls or artificially limit capability.
Your Hub, Anywhere
Use your own machine as the hub — connect from any device, run any API, build natively. No cloud rental. No middleman. Your workstation is the platform.
One app. Everything built in.
No plugins. No extensions. No configuration hell. Install it and everything works.
AI Conversations
Full chat interface with markdown, code highlighting, file attachments, and multi-turn conversations. Connect any LLM provider.
Built-in Browser
Browse the web within the app. Research, pull context, reference documentation — all without switching windows.
Specialist Agents
Expert presets that transform the AI into domain specialists. Frontend, backend, security, DevOps, game dev, and more.
Not one AI. Many experts.
Each agent is a behavioral configuration that shapes the LLM into a domain-specific specialist. Select one and the AI adapts instantly.
TypeScript, CSS, UI/UX, responsive design, PWAs, Core Web Vitals, component architecture.
Node.js, Python, Go, Rust, APIs, REST, GraphQL, databases, microservices, serverless.
OWASP Top 10, XSS, CSRF, SQLi, JWT, OAuth 2.0, encryption, vulnerability assessment.
Docker, Kubernetes, CI/CD, cloud platforms, Terraform, monitoring, load balancing.
PyTorch, TensorFlow, LLMs, fine-tuning, RAG, embeddings, local model deployment.
Unity, Unreal, Godot, game mechanics, physics, multiplayer, C#, GDScript.
Build agents that think like you.
Every built-in agent is just a starting point. Create your own specialist agents from scratch, fork existing ones and customize them, or let the AI generate an agent tailored to your exact workflow. Define behavior rules, tool access, memory scope, and domain expertise — then save and share them.
Create from Scratch
Define a new agent with custom instructions, personality, tool permissions, and domain focus. Full control over every parameter.
Fork & Customize
Take any built-in agent as a template. Adjust its behavior, add rules, restrict tools, or expand its expertise to fit your exact needs.
AI-Generated Agents
Describe what you need in plain language and THERION builds the agent for you — behavioral rules, tool configuration, memory setup, all wired correctly.
Three steps. No friction.
Download & Install
Grab the release for Windows, macOS, or Linux. Double-click. Done. No account required.
Add an API Key
Open Settings. Connect Ollama (free, local), Anthropic, OpenRouter, or any compatible provider. You pay the provider directly.
Pick an Agent
Select a specialist from the home screen. Start a conversation. The AI adapts to the domain instantly.
Built with craft.
Native desktop performance. Web technology flexibility. Everything runs locally on your machine.
Your data stays on your machine. No telemetry. No tracking. No cloud dependency. The backend and frontend ship as a single native binary. Portable. Sovereign. Yours.
Stop renting your AI.
Download THERION. Inspect the code. Use it for free. If you like what you see, tell someone.