Autonomous web research agent. Runs overnight. Handles Cloudflare, dynamic JS, login walls, and rate limits. Built because every other option failed the moment it touched anything real.
Blackreach is an autonomous browser agent built for research tasks that actually need to finish. ReAct loop, custom DOM walker, stealth Playwright. Give it a goal and it runs until done. Overnight, unsupervised, on real sites that push back.
Thought: I need the inscription table on this page
Action: navigate("https://sigla.phis.me/")
Observation: Page loaded. Nav: [About, Database, Signs].
Main: table, 847 rows, columns [ID, Site, Text, Image].
Interactive: pagination controls, export button.
Thought: extract all rows and handle pagination
Action: extract_table(selector=".inscription-table", paginate=True)
...
The DOM walker compresses a 200k token page down to a 2k token observation.
The model sees structure, not noise. Every interactive element gets a numeric ID
so the agent clicks [15], not a CSS selector it has to guess.
Task input
|
v
[ ReAct Loop ]
[ Think (LLM) <---> DOM Walker ]
[ | ]
[ v ]
[ Act ------> Stealth Playwright ]
|
v
Output + task log
pip install blackreach playwright install chromium blackreach run "your task here"
Open source at gitlab.com/null.phnix/blackreach. Issues and PRs welcome.