What if the run stopped halfway?

You asked for more than the daily budget. Ask what actually completed rather than what was planned, size the next run smaller, and pick up the next day.

What if a workflow has been green for six weeks and is doing nothing?

The session needs reconnecting. Empty results are not errors in n8n. Alert on absence, not on errors. fewer than N writes in forty-eight working hours pages somebody.

What if two workflows share an account?

Each is well behaved and together they are not. One writer per account, always. In a multi-account setup, log which account each run used. it is the first thing you want when debugging and it is never there.

What if the stateful sidecar drifted?

The table you added to n8n and the actual state got out of step, and now the workflow is confidently doing the wrong thing. There is nothing to fix inside the workflow. The job is the wrong shape, and the state needs to move to a file the script owns.

How many LinkedIn actions does cli or a workflow tool? use up?

Profile views: 120 a day, reset at midnight UTC. Take a third and you get about forty per run. A four-hundred-name enrichment is a ten-run job at forty a run — a fortnight daily, or a quarter weekly. Compress it and the run stops partway, which is worse than not starting: a partial answer with no clear record of what completed. This is where the sidecar-store instinct comes from, and it does not save you, because the state you needed was "which rows are done", not "which rows were attempted".

CLI or a workflow tool? on LinkedIn, using one prompt

Last updated 2026-09-01 · Tested on a real LinkedIn account

Short answer

Know when to reach for a workflow tool and when the CLI is already enough. You ask in plain English. Your AI assistant runs the LinkedIn CLI for you, so you never type a command yourself.

How many LinkedIn actions does this use?

LinkedIn limits how much any account can do in a day. These are the real numbers, and your assistant checks what is left before it starts.

Profile views: 120 a day, reset at midnight UTC. Take a third and you get about forty per run. A four-hundred-name enrichment is a ten-run job at forty a run — a fortnight daily, or a quarter weekly. Compress it and the run stops partway, which is worse than not starting: a partial answer with no clear record of what completed. This is where the sidecar-store instinct comes from, and it does not save you, because the state you needed was "which rows are done", not "which rows were attempted".
Connection requests: 90 a week, reset Monday. Treat sixty as the working target so a busy week never pushes into a hold. Across five working days that is twelve a day, split into three or four per run at gaps that vary. Written as a self-paced loop in a chat client, not a cron.
Messages: 150 a day. Comments: 100 a day. Reactions: 120 a day. Reading the inbox is unlimited; sending is not. Every write class has the same shape — take a third as the working budget, drip across the working window, hard stop below the cap.
Searches: 1,000 calls a day, per search type. Generous, though every call is billed per page fetched.

Watch it instead

4 minutes

Common questions

What if the run stopped halfway?: You asked for more than the daily budget. Ask what actually completed rather than what was planned, size the next run smaller, and pick up the next day.
What if a workflow has been green for six weeks and is doing nothing?: The session needs reconnecting. Empty results are not errors in n8n. Alert on absence, not on errors. fewer than N writes in forty-eight working hours pages somebody.
What if two workflows share an account?: Each is well behaved and together they are not. One writer per account, always. In a multi-account setup, log which account each run used. it is the first thing you want when debugging and it is never there.
What if the stateful sidecar drifted?: The table you added to n8n and the actual state got out of step, and now the workflow is confidently doing the wrong thing. There is nothing to fix inside the workflow. The job is the wrong shape, and the state needs to move to a file the script owns.
How many LinkedIn actions does cli or a workflow tool? use up?: Profile views: 120 a day, reset at midnight UTC. Take a third and you get about forty per run. A four-hundred-name enrichment is a ten-run job at forty a run — a fortnight daily, or a quarter weekly. Compress it and the run stops partway, which is worse than not starting: a partial answer with no clear record of what completed. This is where the sidecar-store instinct comes from, and it does not save you, because the state you needed was "which rows are done", not "which rows were attempted".

How do I set this up?

Once, and it takes about four minutes. After this you only write sentences.

your terminal

$ npm install -g @connectsafely/cli$ export CONNECTSAFELY_API_KEY=...$ csly list-linkedin-accounts$ export CONNECTSAFELY_ACCOUNT_ID=...$ csly skill install

The package is @connectsafely/cli on npm. You need Node 20 or newer.

Then start your chat with this line, so your assistant uses the CLI instead of trying to open LinkedIn itself and hitting a login page:

“Use the ConnectSafely CLI (csly) for anything that touches LinkedIn. Run csly commands to see what exists, and csly schema before calling one. Don't guess a command name.”

Other things you can do

Stop double-messaging people Stop two of your accounts contacting the same person.Sync your CRM overnight Keep your CRM and LinkedIn in step overnight, every night.Send the client report Produce the client report from real numbers, every month, without touching it.Build your own agent Build an agent that runs your LinkedIn account while you sleep.

Try this on your own LinkedIn account

From $10 a month, on the LinkedIn account you already use, with zero ban risk.

Get your API key Read API Docs