Ask three search engines, not one.
Brave, Tavily, and Exa run concurrently. URLs are normalized, duplicates become agreement signals, and one deterministic ranked list comes back.
$ webctx search "durable agent runtimes"Search across Brave, Tavily, and Exa. Read the part of a URL an agent actually needs. Map sites and explore structured data, all from one small binary.
npm i -g webctxwebctx search "agent web research"01 / Why federated search
Independent agreement is a stronger relevance signal than any provider's confidence score.
webctx treats provider disagreement as useful information. It asks three systems the same question, normalizes their URLs, and rewards the pages they independently surface.
The result is not three pasted lists. It is one stable, deduplicated answer an agent can inspect, save, diff, and cite.
02 / Web discovery and reading
Small enough to learn from --help. Broad enough for an entire research loop.
Search, read, and map the web with plain-text output. For structured records, use the separateresearch, inspect, and execute commands.
Brave, Tavily, and Exa run concurrently. URLs are normalized, duplicates become agreement signals, and one deterministic ranked list comes back.
$ webctx search "durable agent runtimes"A repository becomes an overview. A line anchor becomes those lines. A PR thread becomes that thread. An Actions job becomes that job and log. Normal pages still come back as clean text.
$ webctx read-link <url>Map a documentation or product site into a URL inventory, then let the agent choose which pages are worth opening.
$ webctx map-site <url>03 / The URL already knows a lot
Keep the meaning of the link instead of flattening every page into the same scrape.
A GitHub line link means source lines. An Issue link means a conversation. A review-thread link means one thread. A job link means one CI job. webctx recognizes those shapes first, then uses markdown or page extraction only when the URL does not already describe a cleaner resource.
See what read-link understandsDoes the URL identify something specific?
Return that resource directlySource · conversations · diffs · history · CI jobsIs clean text already available?
Use the markdown pageFast · focused · low noiseDoes the page need rendering?
Crawl it as a fallbackRendered page → useful text04 / The research loop
Terminal-native context means every output is already composable.
Search broadly or map the site first.
Pass the useful URL to read-link.
Keep line, comment, review, diff, or job URLs when you have them.
Pipe, save, quote, or hand the markdown to an agent.
05 / Documentation
Start with short, task-focused guides. The Reference section holds flags, rate limits, GitHub URL formats, ranking rules, and advanced data details.
Browse every guideA few commands to get useful context into your agent.
One query, one deduplicated list of useful links.
Paste a link and let webctx choose how to read it.
Explore a documentation site without flooding an agent with URLs.
Find a source, narrow the URL, and pass clean context to your agent.
Use Firecrawl Alexandria for structured data beyond webpages.
Connect an MCP client to the hosted webctx retrieval engine.
Environment variables, Keychain storage, and multiple independent accounts.
How webctx shares capacity across keys and concurrent coding agents.