Skip to main content
The context-dev CLI runs Context.dev operations from your terminal: brand lookups, crawls, web searches, batches, monitors, and more.
The CLI is generated from the API spec, so command names and flags can change between releases. Treat --help from your installed version as the source of truth.

Install

Download the archive for your operating system and architecture from the latest GitHub release. Extract context-dev, move it to a directory on your PATH, and check it:
In CI, pin a release and verify its published checksums instead of downloading latest on every run.

Set your API key

Using a coding agent? It can create and store a key through agent registration while you confirm in your browser; see the Agent quickstart. The CLI reads CONTEXT_DEV_API_KEY from the environment. Copy a key from the API keys page and export it in the current shell:
In CI, load the key from a secret manager. Avoid the global --api-key flag, because command-line arguments can appear in shell history and process listings.

Make your first request

--type picks the lookup, as in POST /brand/retrieve, and each type takes its own identifier flag:
The command prints the company name. Remove --transform to see the full response. Other lookup types work the same way:

Find a command

The command groups are parse, web, brand, industry, utility, monitors, batch, webhooks:deliveries, people, news, and logs. The CLI has no Scrape, Map URLs, or feedback command yet. Call POST /web/scrape, GET /web/urls, and POST /feedback directly or through an SDK. Common operations:
Flags that take an object, such as --identifier and --timeout-opts, accept a JSON value.

Control output

Use --format json for stable output in scripts, and --transform with a GJSON path to keep only the fields you need, which also keeps large responses out of an agent’s context window:

Use with an agent

Add a short instruction to your repository’s agent file:
AGENTS.md

Troubleshoot