Create local AI agents for your business — without paid APIs, without cloud lock-in, and without touching the terminal.
Status: 🚧 Alpha — the core features are built and run locally from source, but the project is not production-ready: there's no signed installer yet, and you currently run it from source. Expect rough edges. See the roadmap.
License: AGPL-3.0 · Contributing: CONTRIBUTING.md · Security: SECURITY.md
Evano Studio is a free, open-source, local-first desktop application that lets normal people — creators, startups, and small businesses — create, run, and manage local AI agents on their own computer.
It is designed to feel like a simple desktop control center, not a developer tool. You don't install Python packages by hand, you don't memorize Ollama commands, and you don't paste API keys from a cloud dashboard. You open the app and build AI "employees" for your business.
The product is built on three promises:
- Free by default — no paid APIs, no cloud subscriptions, no metered usage.
- Local-first — your models, documents, and data live on your machine.
- Privacy-first — nothing is uploaded to the cloud unless you explicitly choose to.
- Non-technical users who want AI helpers without learning to code.
- Creators who want a private assistant for writing, research, and image prompts.
- Startups and small businesses that want local AI agents for repeatable work (drafting documents, answering from a knowledge base, generating image prompts, running routines).
- Developers and contributors who want a clean, maintainable, open-source AI desktop platform to build on.
Evano Studio does not sell "terminal commands" or "Ollama setup." It packages a professional experience around free, local, open tools so that anyone can run real AI agents without recurring cost or cloud dependency.
Read the full philosophy in docs/FREE_BY_DEFAULT.md.
Evano Studio is a friendly control center for OpenClaw agents:
- Agents — create OpenClaw agents from templates (name, emoji, mission) without touching the terminal.
- Chats — every conversation in one place (from the app, Discord, team runs…) and continue any of them.
- Teams — multi-agent relay workflows: each agent does its part and hands the work to the next.
- Org Chart — a real chain of command: managers can delegate work to their direct reports (enforced as OpenClaw permissions, not just a drawing).
- Office — watch your agents work live: at their desk when busy, in the break room when idle.
- AFM (Agent File Management) — one tidy folder for everything your agents create (
Agents/<name>,Teams/<team>/<member>). - Discord bots — connect each agent to Discord as its own bot with a guided, non-technical flow.
- Knowledge base (RAG) — import documents locally (ChromaDB); agents use them in chats and on Discord (memory sync).
- Documents — every file your agents write, grouped by agent, with preview.
- Local image generation — agents craft prompts and generate images via a local ComfyUI instance.
- Routines & calendar — schedule single agents or whole team workflows.
- Logs & support bundle — diagnose problems with privacy-respecting logs.
Full scope and boundaries are in docs/ROADMAP.md.
| Layer | Technology |
|---|---|
| Desktop shell | Electron |
| Desktop UI | React + TypeScript |
| Local backend | Python + FastAPI |
| Local AI runtime | Ollama |
| Recommended model | Gemma (3/4 class) — falls back to other supported models |
| Other models | Qwen, Llama, Mistral, DeepSeek distill |
| App database | SQLite |
| Knowledge base / RAG | ChromaDB (local persistent, no cloud) |
| Image generation | ComfyUI (optional, local, external service) |
| Website | Next.js + React + TypeScript |
See docs/ARCHITECTURE.md for the full picture.
A pnpm monorepo:
evano-studio/
├── apps/
│ ├── desktop/ # Electron + React + TypeScript desktop app
│ └── website/ # Next.js website (evanostudio.com)
├── services/
│ └── agent-engine/ # Python FastAPI local backend
├── packages/
│ ├── shared/ # Shared types/config
│ ├── ui/ # Shared UI components (if useful)
│ └── config/ # Shared config constants
├── scripts/ # Setup, dev, and health-check scripts
├── data/ # Local-only runtime data (git-ignored contents)
│ ├── sqlite/ # Local dev database
│ ├── chroma/ # Local ChromaDB persistent storage
│ ├── workspace/ # Generated documents/images
│ └── logs/ # Local logs
└── docs/ # Project documentation
See docs/PROJECT_STRUCTURE.md for what each folder is for and how to add modules.
Evano Studio is free and local; a couple of local tools are installed by you — the app guides you and never installs software silently:
- Ollama (required) — the local AI runtime. Install from ollama.com, then install a model from the app's Models page. See docs/OLLAMA.md.
- ComfyUI (optional) — only for local image generation; you install and run it yourself. See docs/COMFYUI_INTEGRATION.md.
- ChromaDB (bundled) — the local knowledge base is a Python dependency of the backend; no separate install. See docs/SQLITE.md and docs/RAG_WITH_CHROMADB.md.
Prerequisites: Node.js ≥ 20, pnpm ≥ 9, Python ≥ 3.10.
# 1. Install JS workspace dependencies
pnpm install
# 2. Backend (local FastAPI agent engine)
cd services/agent-engine
python3 -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
uvicorn app.main:app --host 127.0.0.1 --port 8765 # http://127.0.0.1:8765/docs
# 3. Desktop app (in another terminal, from the repo root)
pnpm dev:desktop
# Optional: the website
pnpm dev:website # http://localhost:3000Then in the app: install Ollama + a model (Models page), create an agent, and chat.
Useful commands
pnpm dev:desktop # run the desktop app (Electron + React)
pnpm dev:website # run the website (Next.js)
pnpm --filter @evano/desktop typecheck
pnpm --filter @evano/website typecheck
# backend tests:
cd services/agent-engine && source .venv/bin/activate && pytestNew here? Start with docs/PROJECT_STRUCTURE.md and docs/ARCHITECTURE.md.
| Document | Purpose |
|---|---|
| Start at the docs index: docs/README.md. Key documents: |
| Document | Purpose |
|---|---|
| docs/PROJECT_STRUCTURE.md | Repository layout and how to add modules |
| docs/ARCHITECTURE.md | System and monorepo architecture |
| docs/DESKTOP_APP.md | The Electron desktop app |
| docs/AGENT_ENGINE.md | The local Python/FastAPI backend |
| docs/OLLAMA.md | Local AI runtime & models |
| docs/SQLITE.md | Local database |
| docs/RAG_WITH_CHROMADB.md | Local knowledge base (ChromaDB) |
| docs/COMFYUI_INTEGRATION.md | Local image generation |
| docs/ROUTINES_AND_CALENDAR.md | Routines & calendar |
| docs/DOCUMENTS.md | Documents & workspace |
| docs/EASY_START.md | Onboarding wizard & setup status |
| docs/TOOLS.md | Safe local tools & agent tool calling |
| docs/DISCORD.md | Optional Discord channel for agents |
| docs/OPENCLAW.md | Evano as an easy front-end for OpenClaw |
| docs/TROUBLESHOOTING.md | Common issues & support bundle |
| docs/FREE_BY_DEFAULT.md | Free/local philosophy |
| docs/SECURITY.md | Security model |
| docs/PACKAGING.md | Desktop packaging |
| docs/ROADMAP.md | Scope and phases |
Evano Studio is in public alpha.
- The core features are built and working locally from source — not a mockup.
- There is no signed installer yet; you run it from source for now.
- It is not production-ready — expect rough edges, bugs, and changes.
- Results and speed depend on your hardware and the local models you choose.
Contributions are welcome. Start with CONTRIBUTING.md and the engineering standards in docs/DEVELOPMENT_RULES.md. Use the issue templates for bugs and feature requests.
Please report vulnerabilities privately — see SECURITY.md. The local-first security model is documented in docs/SECURITY.md.
Evano Studio is open source under the GNU AGPL-3.0 License. By contributing, you agree your contributions are licensed under AGPL-3.0. In short: you can use, study, modify, and share this software freely — but if you offer a modified version as a service, you must share your changes under the same license.