- Home
- AI Computer Use Agents
- Pan-Agent
Pan-Agent
What it is
Pan-Agent is a free, open-source AI assistant designed to run on a user’s own machine. It can capture screenshots, interpret what is on screen, type, click, manage windows, run commands, and browse the web. The project lists more than 20 built-in tools, including terminal, filesystem, browser, web search, keyboard, mouse, vision, memory, and cron. It connects to OpenAI-compatible providers such as OpenAI, Anthropic, OpenRouter, Groq, Ollama, LM Studio, vLLM, and llama.cpp. Scheduled agent runs can be set with cron expressions, while separate profiles keep API keys, model settings, persona, memory, and skills apart. Built-in Telegram, Discord, and Slack bots do not require a public URL. Pan-Agent can also run without a desktop interface as a local HTTP API on localhost:8642, documented as having 50 endpoints. Its 103-pattern approval system classifies commands and blocks catastrophic ones by default. The project says it has no telemetry. The API has no authentication and is intended for a single-user desktop app; Linux control needs X11 or XWayland. Windows installers are unsigned, and the macOS app is not notarized, so first launch may need manual steps.
For and against
For
- Runs on the user’s own machine
- Supports a range of OpenAI-compatible providers
- More than 20 built-in tools
- Catastrophic commands are blocked by default
- Telegram, Discord, and Slack bots need no public URL
Against
- HTTP API has no authentication
- Linux control requires X11 or XWayland
- Windows installers are unsigned
- macOS app is not notarized
The verdict
Pan-Agent combines desktop control, integrations, and a local API with explicit command approval controls. Its unauthenticated API and platform-specific setup caveats matter when deciding where and how to run it.
Who it suits
Pan-Agent suits people seeking control over a locally run AI assistant without vendor lock-in. Its command approvals, profiles, and messaging bots may also fit users who want agent actions and scheduled runs under their own setup.
On the record
- API
- The headless HTTP API binds to localhost at port 8642 and is scriptable from any language.github.com · read 1 Oct 2026
- API security
- The HTTP API binds to 127.0.0.1 and has no authentication; the project says defending against arbitrary local code execution is out of scope.github.com · read 30 Sep 2026
- Approval security
- A 103-pattern system classifies commands as Safe, Dangerous, or Catastrophic; catastrophic commands are blocked by default.github.com · read 1 Oct 2026
- Automation
- Cron expressions can trigger scheduled agent runs such as daily briefings and periodic checks.github.com · read 1 Oct 2026
- Auto-update
- The Tauri updater verifies Ed25519 signatures before installing updates.github.com · read 1 Oct 2026
- License
- Pan-Agent is open source under the MIT license.github.com · read 1 Oct 2026
- License and audience
- Pan-Agent is MIT-licensed and describes itself as built for people who want control of their AI agent without vendor lock-in.github.com · read 30 Sep 2026
- LLM integrations
- It supports OpenAI-compatible providers including OpenAI, Anthropic, Regolo, OpenRouter, Groq, Ollama, LM Studio, vLLM, and llama.cpp.github.com · read 1 Oct 2026
- LLM support
- It supports OpenAI-compatible providers including OpenAI, Anthropic, OpenRouter, Groq, Ollama, LM Studio, vLLM, and llama.cpp.github.com · read 30 Sep 2026
- Memory and profiles
- Persistent memory and isolated profiles store separate API keys, model configuration, persona, memory, and skills.github.com · read 1 Oct 2026
- Messaging integrations
- Built-in bots connect to Telegram, Discord, and Slack.github.com · read 1 Oct 2026
- PC control
- It can take screenshots, type, click, manage windows, run commands, and browse the web.github.com · read 1 Oct 2026
- Platform limitation
- PC-control tools require X11 or XWayland on Linux; pure Wayland is unsupported.github.com · read 1 Oct 2026
- Platform security caveats
- The project says Windows installers are unsigned and the macOS app is not notarized, so first launch may require manual steps.github.com · read 30 Sep 2026
- Privacy
- The project states that it has no telemetry and zero phone-home behavior.github.com · read 1 Oct 2026
- Product
- Pan-Agent is an open-source, self-hosted AI assistant that runs on the user’s own machine.github.com · read 30 Sep 2026
- Safety controls
- A 103-pattern approval system asks before dangerous commands and blocks catastrophic commands by default.github.com · read 30 Sep 2026
- Security limitation
- The HTTP API has no authentication and is intended for a single-user desktop app.github.com · read 1 Oct 2026
