Documentation

Find the page. Read the useful part. Hand it to an agent.

Start with a working example and go only as deep as the task requires. The Guides focus on outcomes; the How it works pages explain the two technical ideas that make webctx useful: automatic read fast paths before crawling, and cross-provider search ranking.

Start

Quickstart

Install webctx, add the keys you need, and get useful web context in a few commands.

Install

Install webctx from npm, a release binary, or source.

Credentials

Add only the provider keys needed for the webctx commands you use.

Guides

Search the web

Search Brave, Tavily, and Exa together and get one clean list of useful pages.

Read a URL

Read any useful URL as clean text; webctx automatically chooses native, direct-markdown, or crawl paths.

Map a site

Discover the useful URLs on a site before deciding which pages to read.

Use webctx with agents

Practical copy-paste workflows for research, repository understanding, code review, CI debugging, and documentation work.

How it works

How URL reading works

See how webctx chooses the cleanest way to read a URL before falling back to a full page crawl.

How search ranking works

See how webctx combines Brave, Tavily, and Exa into one deterministic result list.

Reference

CLI reference

The small command and flag reference for everyday webctx use.

Troubleshooting

Quick fixes for missing keys, GitHub permissions, rate limits, Firecrawl failures, and installation problems.

Changelog

Release notes for webctx.