Find
A command away
Search by name, phone, email or username, and read the candidates right in your terminal.
Find
Search by name, phone, email or username, and read the candidates right in your terminal.
Script
Add --json for the raw response. Results go to stdout and progress to stderr, with predictable exit codes.
Queue
Queue a lookup, come back for the result, and replay its progress whenever you need it.
Use Node.js 20.9 or later and npm. Sign in with an API key, or try the sandbox first without an account.
npm install -g @reloadapp/deepsearch
deepsearch --version
deepsearch helpAgents that can run terminal commands can use the CLI too, including Claude Code, Codex, Cursor and GitHub Copilot.
Create an API key in your developer settings. It draws on the same usage as the app.
Save it with deepsearch login, which writes a file only you can read, or pass it through DEEPSEARCH_API_KEY.
Run deepsearch status to confirm the endpoint and features before your first lookup.
Search by any identifier, build a sourced report, and ask a question grounded in it.
searchdossierchatQueue a lookup as a job, then fetch it or replay its progress later.
jobs createjobs listjobs getjobs eventsStore a key, inspect your configuration and check the API's status.
loginconfigstatusdeepsearch search "Ada Lovelace"
deepsearch search ada@example.com --type email --json
deepsearch search ada --type username --platforms linkedin,githubNode.js 20.9 or later and npm. Install it globally with npm install -g @reloadapp/deepsearch, or run any command once with npx.
Yes, as @reloadapp/deepsearch. It installs a single deepsearch command and has no runtime dependencies.
With a DeepSearch API key. It checks the --api-key flag first, then the DEEPSEARCH_API_KEY environment variable, then the file deepsearch login saves in ~/.deepsearch.
Yes. Pass it through DEEPSEARCH_API_KEY from your secret manager, or with --api-key for a single call. Nothing is written to disk unless you run deepsearch login.
Add --json to print the raw response on stdout. Progress and usage go to stderr, so piping stays clean. It exits 0 on success, 1 on an API error, 2 on a usage error and 130 when interrupted, and it honours NO_COLOR.
Yes. Add --sandbox and the public key dsk_test_deepsearch to any lookup. You get the same fixed Ada Lovelace results every time, unmetered.
Search, dossier and chat. Create a job, then fetch it with jobs get or replay its saved progress with jobs events. Add --idempotency-key so a retry never runs the same lookup twice.
Revoke the key in your developer settings and it stops working everywhere. To clear the saved copy, delete the file deepsearch config path prints.