Skip to content
contextretrieval-netPublic

About

Context Retrieval CLI: live web pages as clean Markdown, text or HTML for AI agents. Commands follow the live API spec.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

5 Commits

Folders and files

Repository files navigation

Context Retrieval CLI

What happens after the fetch: chunking, retrieval and citations for web RAG.

contextretrieval fetches live web pages as clean Markdown, text or HTML, straight from your terminal or your agent's shell. It is the command line client for the Live Web Data API by Context Retrieval.

Quick start

npx @contextretrieval-net/cli login                                   # paste your API key (not echoed)
npx @contextretrieval-net/cli fetch https://example.com -o text       # the page as Markdown

Get a key at contextretrieval.net. Node.js 18 or later.

Install

npm install -g @contextretrieval-net/cli
contextretrieval --help

Use

contextretrieval fetch https://example.com/a https://example.com/b      # up to 10 URLs; JSON with results[] and statuses[]
contextretrieval fetch https://example.com -o text                      # just the content (Markdown, else text, else HTML)
contextretrieval fetch https://example.com -o html --html               # the page HTML
contextretrieval fetch https://example.com --country-code de --render   # from Germany, rendered in a real browser
contextretrieval fetch --data @request.json                             # the whole request body from a file (- for stdin)
contextretrieval account                                                # plan, credits and rate limit
contextretrieval fetch --help                                           # every option, with the API's own descriptions

A fetch that is still running after about 100 seconds is collected automatically with the same Idempotency-Key, so it is charged once.

The commands follow the live API

contextretrieval builds its commands and options from the API's OpenAPI document at https://livewebdata.com/openapi.json, cached for an hour. When the API gains a field or an endpoint, the option or command is there on your next run, without updating the CLI. Offline, it uses the last cached copy, then the copy bundled with the installed version (contextretrieval spec shows which one; --offline skips the download). An hourly job here publishes a new version whenever the API spec changed.

Configuration

Setting Option Environment Default
API key --api-key LIVE_WEB_DATA_API_KEY the key saved by contextretrieval login (~/.config/contextretrieval/config.json, mode 0600)
API base URL --base-url LIVE_WEB_DATA_API_BASE_URL https://livewebdata.com/api
API spec URL --spec-url LIVE_WEB_DATA_SPEC_URL https://livewebdata.com/openapi.json
Output -o json|text|<field> json
Update notice CONTEXTRETRIEVAL_NO_UPDATE_NOTIFIER=1 turns it off once a day, terminals only

Requests and your key only ever go to the base URL, never to a server named in the spec.

Exit codes

Code Meaning
0 OK
1 The API answered with an error (JSON on stderr, with its requestId), every URL failed with -o text, or the fetch was still running at --max-wait
2 Usage: unknown command or option, a value the API does not accept
3 Network: no answer. The error carries the Idempotency-Key to collect the result without a new charge

Links

License

MIT, © Teneo Labs Inc.

About

Context Retrieval CLI: live web pages as clean Markdown, text or HTML for AI agents. Commands follow the live API spec.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages