Outcome · Data
Data agents for LinkedIn
Get LinkedIn out of the browser and into something you can query: a file, a CRM record, a model's context window.
How these act
- Your LinkedIn session — Acts as you, from your connected account, under pacing limits.
You bring
- A LinkedIn account the reads run through, and an API key if you are calling the MCP endpoint.
- Somewhere for the data to land — a file, a sheet, a CRM, or your own code.
- A use that respects the profile owner: this reads data, it does not license it.
What it takes
The capabilities this job runs on
A capability is one thing an agent can do, and each one names what performs it: a live MCP tool, a skill, or a workflow. 8 of 11 can be performed today.
Check the remaining LinkedIn budget
Low riskHow many LinkedIn tool calls this workspace has made this month, broken down by tool, plus the rate limit in force. The call an agent makes to pace itself instead of discovering the limit by hitting it.
Runs on get_usage · Read only
Find people matching an ICP
Medium riskRuns an ICP description through a connected account's LinkedIn search and returns up to 50 profiles per call. Profiles LinkedIn masks as "LinkedIn Member" because they are out of network are dropped rather than returned with a blank name. Typed Sales Navigator filters are used only when the account has a Navigator seat; otherwise every criterion is folded into one keyword string.
Runs on search_linkedin_people · Your LinkedIn session
Map a company from its people
Low riskBuilds a picture of a company out of the employees a search returns: functions, seniority spread, recent joiners and recent leavers. It is a sample of who is visible to the searching account, and the skill is required to label it as one rather than to present it as headcount.
Performed by a skill — instructions your AI client follows. · Read only
One company, mapped from its people
Medium riskSearch the company, cluster what comes back by function and seniority, sample a few profiles for tenure and job-change dates, and write the map with its sampling limits stated in the document itself.
Performed by a workflow — a sequence of steps, not a single call. · Your LinkedIn session
Read one profile in depth
Medium riskOne call returns identity, current role and the date it started, follower count, the premium flag and the latest post. Every enrichment field is best-effort: a profile that hides its activity still returns its identity, with the post fields simply absent.
Runs on get_linkedin_profile · Your LinkedIn session
Read what someone has posted
Medium riskUp to twenty recent posts by one member, newest first, with reposts flagged. An empty array — not an error — is the normal answer for a senior buyer who does not post, and a workflow has to handle that as a case rather than as a failure.
Runs on get_linkedin_posts · Your LinkedIn session
See which LinkedIn identity is available
Low riskThe accounts this workspace can act as, each with status, warm-up stage, how much of today's invite and message budget is spent, and a health score. Reads our own database, so it costs nothing against the LinkedIn rate limit and is the cheapest first call an agent can make.
Runs on list_sender_accounts · Read only
Write a brief on one person
Low riskCompresses a profile and its posts into a one-screen brief in which every claim names the call that produced it. Anything the tools did not return is written as "not visible" rather than inferred, because an invented detail in a first message is worse than a missing one.
Performed by a skill — instructions your AI client follows. · Read only
Not built yet
These are declared in the registry so a product page can say “not yet” instead of implying it. Nothing performs them today:
- Read a company's open roles — Not built yet. Hiring signals available today are indirect: recent joiners and job-change dates on individual profiles. The job board itself is not readable through any tool.
- Read a LinkedIn company page — Not built yet. No tool reads a company page, so headcount, follower count and the company's own description are not available to any product here. Company research is done from the people a search returns, which is a sample, not a roster.
- Write a researched lead into a CRM — Not built yet. Research comes out as files you own. There is no CRM write path, so the last step is still a copy-paste or an import.
The products
1 product does this job
Each card is a real catalogue row: its price, its risk level and how it acts are the fields the product itself declares, not marketing written per page.
What it costs
Only the tiers this outcome actually uses
- Pro kit$491 product
- A skill plus MCP wiring, workflows and the files they read.
There is no payment processor connected yet. Checkout records a real order against your workspace and tells you exactly what happens next — nothing is charged, and no card is asked for. The whole ladder is on pricing.
What it risks
The part most sites leave out
This is reading at human pace through one account, not a scrape. There is a hard rate limit per workspace, and it is there so that an agent stuck in a retry loop hits us instead of hitting LinkedIn.
- Medium risk
- Acts through your own session or writes with a human approving each action. Restriction is unlikely but not impossible.
LinkedIn does not endorse, certify or sanction any of this, and nobody — including us — can promise an account will not be restricted. What we can do is keep the irreversible actions behind a person, cap the pace, and tell you which side of that line a product sits on before you buy it. The mechanics are on security.
Other outcomes: Sales · Recruiting · Content · Networking · Operations