Metadata-Version: 2.5
Name: cairn-browser
Version: 0.1.0
Summary: A browser memory for AI agents: learn a site once, replay the trail with no model calls.
Project-URL: Homepage, https://github.com/rohit-jsfreaky/cairn
Project-URL: Repository, https://github.com/rohit-jsfreaky/cairn
Project-URL: Issues, https://github.com/rohit-jsfreaky/cairn/issues
Author: Rohit Kashyap
License-Expression: MIT
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Internet :: WWW/HTTP :: Browsers
Classifier: Topic :: Software Development :: Testing :: Acceptance
Requires-Python: >=3.11
Requires-Dist: playwright>=1.62.0
Requires-Dist: sibyl-memory-client>=0.8.0
Provides-Extra: dev
Requires-Dist: fastapi>=0.141.1; extra == 'dev'
Requires-Dist: pytest>=8.0; extra == 'dev'
Requires-Dist: uvicorn>=0.52.4; extra == 'dev'
Provides-Extra: market
Requires-Dist: fastapi>=0.141.1; extra == 'market'
Requires-Dist: uvicorn>=0.52.4; extra == 'market'
Requires-Dist: x402[evm,fastapi,requests]>=2.21.0; extra == 'market'
Description-Content-Type: text/markdown

# Cairn — a browser memory for AI agents

**Your AI can use websites. But it forgets how, every single time. Cairn makes it remember.**

Your AI walks a website once — signing in, clicking, reading — and Cairn writes down the
route. Every run after that follows the route instead: **one call, no page reading, no model
calls at all.** When the site changes, Cairn repairs the one step that moved and keeps the
rest.

```bash
pip install cairn-browser
playwright install chromium        # the browser is a separate download

cairn run --site github.com --task "count open issues"
cairn show github.com              # the route, step by step
cairn forget --site github.com     # make it forget
```

This is the engine. Most people want it wired into their AI instead, which is
[`cairn-browser-mcp`](https://pypi.org/project/cairn-browser-mcp/) — Cairn as MCP tools for
Claude Code, Cursor or Codex.

**No API key.** Cairn never calls a model; yours does the thinking. Memory is a local SQLite
file and needs no account.

Optional extra, for selling and buying trails between machines over x402 on Base:

```bash
pip install "cairn-browser[market]"
```

Full documentation, the design, and the deletion test that proves the memory is load-bearing:
**https://github.com/rohit-jsfreaky/cairn**

MIT licensed.
