API glossary
The vocabulary you meet when you start automating LinkedIn and messaging channels, defined plainly — account ids, webhooks, shared inboxes, linked numbers and the rest.
You tell the assistant you already use what to do. Your LinkedIn profile or page does it daily, at human pace: safe bulk reactions that reach people you could never target, posts, invites, replies or comments. Everyone it brings writes to you on LinkedIn, WhatsApp, Instagram, email or your website chat, where your team and AI agents close deals and run tasks.
Everything from $9 per month, first 30 days free
The white label is part of the same subscription: your name, your logo and your web address.
“Organically boost my LinkedIn profile and page.”
“Post about what we do twice a week.”
“Invite the heads of sales at Berlin fintechs.”
“Answer everyone who replies.”
You already know how to ask ChatGPT or Claude for something. That is the whole interface — and Gemini, Grok, Cursor and Claude Code work the same way. Nothing to build and nobody to brief.
Real posts, real comments, real invitations, from the account people already know. Paced like a person, because it is your account. Anything with your name on it waits for your yes.
They write from WhatsApp, Instagram, email or your website chat; your team and your AI agents answer in one room, the deal is closed there, and the work carries on in the same thread.
A group of people building LinkedIn and AI apps for a living. Free to join, and there is nothing to buy to be in it.
Tell us which of those you came for. We read every one of these and reply by email — there is no automated sequence behind it.
The MakeHybrid API can carry your name instead of ours, for your own customers, at no extra charge. There is no reseller tier and no revenue share: you pay the same price per seat that everyone else does, and what you charge your customers is your business. Your customers count towards your own tier, so reselling is how most people reach $7.
Your pages
Your product name, logo and color on every page your customers see, on your own web address. Your words for the home page, terms and privacy — or links to pages you already have. Your Stripe keys, so your customers pay you directly.
Your customers' AI
The one-address AI connection, under your name. Your customers paste it into Claude or ChatGPT and the tools are called after your product, not ours. You choose whether it covers everything you offer or only certain areas.
Both, and that is the point
They are not two versions of the same idea — they are different layers, and they stack:
What happens to a LinkedIn account is LinkedIn's decision, and nobody can promise otherwise. What this is built for is never to be the reason.
A campaign takes one action per turn and counts the day's budget itself, from the actions already taken. It waits the same gap whether it found something to do or nothing at all, and when LinkedIn asks for a pause the API answers with how long to wait rather than retrying. A burst is what gets an account checkpointed, and nothing here can burst.
The sign-in happens from a location you choose, because a sign-in from an unexpected country is what usually sets off extra checks. No pods, no shared logins, no account but your own, and only your own organization can reach it. If LinkedIn does put a checkpoint on it, the connection says so, so you stop instead of hammering it.
A content campaign drafts; publishing waits for your yes, and that is what it does unless you change it. A draft waiting on you opens a task in your project room, so you are told rather than having to go and look, and you can edit it, move it to another time or set it aside. Anything with your name on it is something you said yes to first.
Three ways to end up with the same result without writing the code: buy something already running on it, have it built for you, or join the people who build these things and ask them.
Ready-made products, and teams that do the work for you. If one of them is already what you were going to build, buy it instead.
Reacts to your prospects’ posts so they notice you, then greets the ones worth approaching.
The same idea without connecting your account: a company page run for you does the noticing.
Paid Ads managed by specialists, and the integrations and automations built by the same team.
Google Ad Grants for nonprofits, structured and kept within policy so the grant is not lost.
Everything
$9 / seat / mo
One seat is all of these at once, not a choice between them: a LinkedIn account, a WhatsApp number, a LINE account, an X account, and a ChatWoot account with its Instagram, Facebook, TikTok, email, SMS and web chat inboxes. A second LinkedIn account is a second seat. Project rooms, the people in them and their AI agents are unlimited.
Project rooms are free to try without a subscription. What a room does on LinkedIn goes through the plan.
What the words mean, a recipe per job, what we have learned running this, and an example you can run for every operation. All of it in one index.
The vocabulary you meet when you start automating LinkedIn and messaging channels, defined plainly — account ids, webhooks, shared inboxes, linked numbers and the rest.
What we have learned building and running LinkedIn automation — mostly about the parts that go wrong, because those are the parts worth writing down.
Ready-to-copy recipes that each solve one job end to end — connect an account, answer every new message, run a campaign. Copy one, change the account id, and it works.
One page per operation, with a request you can run and the answer you get back, generated from the same catalog the API itself is served from.
No. If you use an AI assistant that supports MCP — Claude and several others do — you paste one address into it and ask for what you want in ordinary English. The developer route is there for people who want it, not instead of this one.
No. On the MakeHybrid API it is forwarded once to complete the sign-in and is never written down, logged, or sent back to you. If you would rather not type it at all, you can connect with a session cookie instead.
Nothing can promise otherwise — it is LinkedIn's decision. What we do about it: the MakeHybrid API signs in from a location you choose, because a sign-in from an unexpected country is what usually triggers extra checks. Use it at a human pace and keep an eye on the dashboard.
A seat is one of each kind of account your team talks through: one LinkedIn account, one WhatsApp number, one LINE account, one X account, one ChatWoot account. You have as many seats as the most you have connected of any one kind, so two LinkedIn accounts and one WhatsApp number are two seats. You are billed per seat, per month — not for calls, rooms, people or AI agents. Disconnect an account and the charge follows. It is $9 a seat up to ninety-nine and $7 each from a hundred, and the cheaper rate applies to all of them: a hundred seats cost $700, not ninety-nine at $9 plus one at $7.
Thirty days, for one seat — one account of each kind: LinkedIn, WhatsApp, LINE, X and ChatWoot — enough to find out whether this does what you need. If you want more before the trial is over, start the subscription from your account page and you are billed per seat from that day.
Yes. Connect as many as you need and name the one you mean on each request. Your assistant can see them all too.
Everything answers on makehybrid.com: the LinkedIn API
under /linkedin, the messaging channels under
/inbox/v1, /webchat/v1, /whatsapp/v1,
/discord/v1, /slack/v1 and
/x/v1. The key
selects your organization. Skip this if you are using an assistant — it does all of
it for you.
LinkedIn API https://makehybrid.com/linkedin
Messaging channels https://makehybrid.com/inbox/v1
https://makehybrid.com/webchat/v1
https://makehybrid.com/whatsapp/v1
https://makehybrid.com/discord/v1
https://makehybrid.com/slack/v1
https://makehybrid.com/x/v1
From an AI https://makehybrid.com/mcp/<your API key>
curl "https://makehybrid.com/linkedin/api/v1/users/me?account_id=…" \
-H "X-API-KEY: lk_…"
Errors are JSON and say what went wrong:
{"type":"errors/…","status":402,"detail":"…"}.
The whole catalog is also published as plain text at
/llms-full.txt and as JSON at
/linkedin/mcp/manifest.