# OpenErrand

> A universal, safe way to run actions in a real browser — for a product automating a flow for its users, or for an individual automating their own. An implementation of OBEP, the open Browser Execution Protocol. Local-first: credentials never leave the device.

## Docs

- [Quickstart](/docs/QUICKSTART.md): The easiest way to add OpenErrand is to let your AI coding tool set it up.
- [Integration Guide — zero to first task](/docs/INTEGRATION.md): This is the developer front door — for a product that wants to run a browser action for
- [Author a playbook with an LLM coding tool](/docs/PLAYBOOK_AUTHORING.md): The fastest way to write a playbook is to let your coding assistant (Claude Code, Cursor,
- [Quickstart: driving a task with an LLM](/docs/LLM_DECIDER.md): OpenErrand is "the pipe" — you bring the intelligence.
- [OBEP Security Model](/docs/SECURITY_MODEL.md): This document states the trust model plainly so a customer's security team can
- [The Open/Closed Boundary (OBEP vs OpenErrand)](/docs/OPEN_VS_CLOSED.md): Governing rule: anything whose secrecy would create a vulnerability MUST be
- [OBEP Error Taxonomy](/docs/ERROR_TAXONOMY.md): Stable error codes carried on status messages (phase: "error", code).
- [Self-Hosting an OBEP Relay](/docs/SELF_HOSTING.md): OBEP is self-hostable: the reference relay in obep/relay-reference
- [Enterprise Deployment (force-install + managed config)](/docs/ENTERPRISE_DEPLOYMENT.md): Force-install via Google Workspace admin policy is the primary B2B distribution
- [OpenErrand — Privacy Policy](/docs/PRIVACY.md): Last updated: 2026-06-15

## Full text

- [All docs, concatenated](/llms-full.txt)