Osaurus

Private AI assistance

Use local models for everyday tasks like writing, coding, or file management without sending data to any server.

Autonomous agent workflows

Build agents that run on voice control, watch folders, or execute parallel jobs while you sleep.

Offline productivity

Work with AI entirely offline—no Wi-Fi required—ideal for secure environments or travel.

File and task management

Drop in folders, assign tasks, and let Osaurus read, write, and execute commands like a local AI assistant.

Cross-model collaboration

Start with local models for privacy, then switch to ChatGPT, Claude, or Gemini for tasks that need cloud-level reasoning.

Development and experimentation

Use the open-source MIT license to fork, modify, or ship your own AI agent builds.

Native Swift for Apple Silicon

The app is built natively for M-series chips, delivering fast inference without round trips to a server.

Local-first model support

Connect Ollama, MLX, or LM Studio in one click—runs fully offline on your own chip.

Persistent memory

A shared memory system lets you carry context across all models, local or cloud.

Autonomous agent execution

Build agents that run independently—voice control, folder watchers, browser plugins, and parallel jobs.

Full offline capability

Works with Wi-Fi off; no telemetry, no training on your data, no servers in the loop.

Cryptographic identity

Offers cryptographic identity for secure, private AI agent operations (implied by the short description).

Open source and free

MIT licensed, no subscription, no usage caps, no billing—yours to use and modify forever.

Drop-in file and task system

Drop a folder, give it a job—reads your files, runs your tools, and remembers your work.