Serious work. Fully local.
Artha is a local-first AI workspace that generates documents, indexes your files, and runs agents — all on your hardware. No cloud calls. No accounts. No telemetry.
An AI workspace that earns trust.
Every capability built around a single rule: your files, prompts, embeddings, and model outputs never leave the machine they were generated on.
Native artifacts
Generate Word, Excel, PowerPoint, and PDF files from natural language. Every artifact stays on your machine.
MCP-native
Any Model Context Protocol server becomes a skill your agent can use. Bring your own integrations.
Local RAG
Index folders with Ollama embeddings. Search, cite, and chat with your files — no cloud.
Ollama-first
Auto-detect installed models, pull new ones, and switch between Llama, Mistral, Qwen, and more.
Zero telemetry
No accounts. No tracking. No background phone-home. The binary is auditable.
Sandboxed tools
Optional Docker sandbox for untrusted tools. Run agents without giving them your shell.
Local-first, honestly priced.
Everything runs on your machine on every plan — your files never leave it. License keys verify offline; no account, no phone-home, ever.
Try the local AI coworker for real — no account, no card.
- Chat, tools, local RAG & memory
- BYOK cloud models (your key)
- 5 generated documents / month
- 1 saved context pack
- Community support
The full solo experience — a fresh offline key on every renewal.
- Everything in Free, uncapped
- Unlimited document generation
- Scheduled tasks & unlimited packs
- Starter skill templates (legal · finance · ops)
- Priority email support
One hub machine, your whole team working with shared context. Min 5 seats.
- Everything in Personal
- LAN team hub — data stays in your office
- Shared memories & shared context packs
- Seat-capped roster + API keys
For regulated teams that must prove what their AI did. Min 5 seats.
- Everything in Team
- Audit-log export — every tool call, attributable
- Org hub deployment + role controls
- Security-questionnaire support
Air-gapped / on-prem Enterprise from $7,500/yr — talk to us.
Before your first launch.
Three things worth knowing.
- 01
Requires Ollama running locally.
We recommend
llama3.2:3b-instruct-q4_K_Mto start — fits in 8 GB of RAM and is fast on most laptops. - 02
macOS builds are signed & notarized by Apple.
They install like any Mac app — no warnings. The Windows installer is unsigned during the beta, so SmartScreen asks once: click “More info” → “Run anyway”. A signed Windows build ships when we leave beta.
- 03
The app runs entirely on your machine.
No telemetry, no account, no cloud — your data stays local.