Skip to main content
Ten minutes, one terminal, a working Octolens workspace at the end. Install the CLI first if you have not (Install & login).
1

Sign up — or log in

No Octolens account yet? The CLI creates the workspace itself:
The browser completes the signup and the CLI ends authenticated; on a box with no browser at all, octolens signup --headless runs the whole flow in the terminal with an emailed code. Both modes — and the Agents plan such workspaces start on — are on Signup & the Agents plan.Already have a workspace? Log in instead:
On a headless box, use an API key — run octolens login --with-key ak_live_xxx or set OCTOLENS_API_KEY; details on Install & login.
2

Run your first search

One command, no setup — a one-time, AI-scored search across the platforms enabled for your workspace:
Results are ranked by relevance and never mix into your mentions feed. On the Agents plan searches draw on a lifetime allowance — Signup & the Agents plan explains the numbers and the upgrade path.
3

Initialize your workspace

New workspace? The guided onboarding builds it from the terminal — company profile, the keywords you want monitored, noise filters and your first feed:
It is safe on an existing workspace too: it reports what is already set up and only offers the missing steps. See Initialize a workspace for the whole flow.
4

Check who you are

Confirm the workspace and the credential’s scope before doing anything else — this is also the preflight your scripts should run:
5

Read your mentions

The ten newest mentions Octolens has collected for you:
Narrow it while you explore — add --source reddit --sentiment negative for the classic “what is bothering people” view.
6

Triage one mention

Take a sourceId from the list you just printed (the examples below use a placeholder), inspect the mention, then act on it:
mentions update corrects the AI’s relevance or sentiment call; mentions engage marks that you have responded to the conversation.
7

Save a feed

A feed is a saved filter over your mention stream. Describe the one you want in plain language and let the AI build the filter:
From now on octolens mentions list --feed 42 (use the id the create printed) reads that slice directly, and octolens feeds watch follows your mentions live in the terminal.
8

Read the numbers

Mention volume over time, as a sparkline plus per-bucket counts:
octolens analytics sentiment and octolens analytics keywords cut the same data by tone and by term.

Scripting what you just did

Everything above works headlessly: add --json and each command emits exactly one JSON document on stdout (Output modes defines all three renderings). Every command also accepts the global flags, generated here from the CLI’s own contract manifest: When scripting, branch on exit codes (or the --json error envelope’s code), never on message text: The full rules live on The —json contract and The error envelope.