docs
browser-memory docs
An open catalog of pre-built web skills for AI agents: reusable, battle-tested recipes for operating specific sites. Your agent installs one with a single command and runs it with the browser it already has — so it stops re-learning every page.
What is browser-memory?
Without a skill, a browsing agent pays the same tax on every run: take a screenshot, snapshot the DOM, guess selectors, retry, and burn tokens re-deriving a page someone else already figured out.
browser-memory removes that tax. Each skill is a SKILL.md — one task on one site, written down with the exact selectors and, whenever possible, the direct in-page request. bmem add registers it as a native skill of your agent, and the agent runs it with its own browser. When a site ships new markup and a skill breaks, it is fixed once in the catalog, for everyone.
The system at a glance
bmem is a thin, consume-only CLI: it finds and installs skills, and never touches a page or a cookie. The skill itself runs in your browser:
The pieces
| Piece | What it is | Where |
|---|---|---|
bmem | Thin, consume-only CLI. Searches the catalog and registers skills into your agent — your agent runs them with its own browser. | npm · GitHub |
| Skills catalog | Hosted, read-only registry of curated skills, one per task per site. Nothing to install. | api.browser-memory.com |
SKILL.md | The unit of memory: one task on one site, as a browser-agnostic recipe with exact selectors and the direct request. | format |
llms.txt | Plain-text index of the whole catalog. Zero auth, zero install — an agent can just fetch it. | /llms.txt |
Which way in?
| If you… | Use |
|---|---|
| run an agent that drives a browser (Playwright MCP, claude-in-chrome, …) and want pre-built skills for it | bmem CLI |
| just want to read what the catalog knows, no install | llms.txt · ecosystem |
| need a site or task the catalog does not cover yet | request it |