How to monitor LinkedIn replies via API
Connect LinkedIn outreach to a custom dashboard with a REST API or MCP server. Fetch replies, classify intent and draft answers without opening LinkedIn.
Get startedShort answer
To monitor LinkedIn replies in a custom dashboard, run outreach through a tool that exposes a REST API or MCP server, then poll that API for new messages and render them in your own interface. You need an API key or OAuth connection, a scheduled fetch, and a way to store the last seen message ID so you do not show the same reply twice.
Why you can't connect LinkedIn directly to a dashboard
If you try to build a dashboard on top of LinkedIn's own API, you hit a wall fast. LinkedIn serves approved partners and specific company use cases, not a founder who wants to pull personal message threads into a custom screen. Most accounts cannot get the scopes you need for messaging.
That leaves two options. You can ask every lead to move to email, which loses the point of LinkedIn outreach. Or you can run outreach through a tool that already holds the LinkedIn connection and exposes a REST API or MCP server. Your dashboard then talks to that tool, not to LinkedIn.
The second option also gives you a stable data shape. A reply comes back with a lead ID, a campaign ID, the message text, the timestamp and an intent label. You do not have to parse LinkedIn's HTML or scrape a page. You call an endpoint and get JSON.
Direct LinkedIn API
- Apply for partner access and wait for review
- Ask for messaging scopes that most accounts cannot get
- Build around a connection you do not control
API layer from your outreach tool
- Call a REST endpoint with your API key
- Read replies as JSON in the same shape every time
- Keep the LinkedIn connection inside the tool that manages it
How the data moves from LinkedIn to your dashboard
A reply dashboard is a reader, not a second inbox. The outreach tool sends the invite or message, LinkedIn delivers the reply, and the tool stores the message with the lead ID, the campaign, the text and the time. Your dashboard asks for what changed since the last fetch.
Keep the flow one way for reads. Your dashboard should not write to LinkedIn on every page load. Fetch on a schedule, store the response, and render from your own database. That keeps the dashboard fast and stops a page refresh from turning into an API call.
Use a cursor, not a page number. New replies arrive at the top, so page 2 changes as soon as a message lands. A cursor such as the last message ID or a timestamp gives you a stable place to resume.
- 1
Outreach tool sends the message
A campaign sends a connection request or a follow-up from the connected LinkedIn account.
- 2
Lead replies on LinkedIn
The reply arrives in the tool's inbox and gets a lead ID, campaign ID and timestamp.
- 3
Tool stores the reply and intent
The tool classifies the reply as interested, question, not now or another status you can filter on.
- 4
Your dashboard fetches new replies
A scheduled job calls the REST API with a last-seen timestamp or message ID.
- 5
Dashboard renders the thread
You show the lead name, profile URL, message text, campaign and intent in your own layout.
How to set up the API connection
Start with a read-only key if the provider offers one. You want to prove the fetch works before you add any write actions. Store the key in an environment variable, not in your front-end code.
Use a last-seen cursor. The first call can ask for recent replies. Every later call sends the newest message ID you already have, so you only get new messages. That avoids duplicates and keeps the response small.
Log the raw response for a day or two. When a field changes, you want to see the payload that broke your parser. A small JSON log in your database is enough.
Paste this job
GET /v1/replies?after=msg_9f2c&limit=25
Authorization: Bearer $OUTREACH_API_KEY
Accept: application/json
200 OK
{
"data": [
{
"message_id": "msg_9f2c",
"lead_id": "lead_1042",
"lead_name": "Maya Chen",
"profile_url": "https://www.linkedin.com/in/maya-chen-example",
"campaign_id": "camp_22",
"text": "Can you send the pricing for a team of 10?",
"intent": "interested",
"received_at": "2025-02-18T09:14:00Z"
}
],
"next_cursor": "msg_9f2c"
}How to draft answers from API replies
Once the reply is in your dashboard, you can pass the thread to an LLM and ask for a draft. Give the model the lead's message, the campaign goal, your tone and any objection-handling notes. Ask for a short reply that answers the question and ends with one clear next step.
Keep the draft separate from the send action. A draft button that fills a text box is safer than a script that posts automatically. If you do post through the API, log the message ID and the user who approved it.
Write the prompt so the model cannot invent a price or a feature. Tell it to use only the facts in the thread and to ask a question when it is missing context.
Example
- Prospect
This looks interesting. How does pricing work for a team of 10?
- Suggested reply
For a team of 10, we have a team plan that includes shared workspaces. I can send the exact breakdown or we can take 15 minutes to see if it fits. Which do you prefer?
- You
For a team of 10, we have a team plan that includes shared workspaces. I can send the exact breakdown or we can take 15 minutes to see if it fits. Which do you prefer?
How to keep the dashboard useful after the first reply
An API feed gets noisy if every reply looks the same. Add a small intent field and a status field to your dashboard. Interested replies go to the top. Questions get a draft. Not now replies get a follow-up date. That is enough to keep a founder or SDR from reading every thread in order.
Use a table to decide what to build first. You do not need a full CRM. You need a list, a thread view and a way to mark a reply as handled.
Keep the first version small. A single page with a filter for interested replies and a text box for a draft will do more than a chart of reply counts. Add analytics after the list is reliable.
Reply list
- Dashboard feature
- Shows new replies with lead, campaign and intent
- What it does
- Yes, before any charts
Thread view
- Dashboard feature
- Shows the full message history for one lead
- What it does
- Yes, because context is the point
Draft button
- Dashboard feature
- Sends the thread to your LLM and fills a text box
- What it does
- Yes, if you review before sending
Analytics
- Dashboard feature
- Counts replies by campaign and week
- What it does
- Later, after the list is reliable
Doing this with Kinlead
Kinlead exposes a REST Agent API and an MCP server. Chat apps connect with OAuth, and scripts use API keys. You can find leads, check replies and draft answers from Claude, ChatGPT, Cursor or another MCP client. The Agent API and MCP page covers the connection details.
Replies from every campaign land in one inbox, classified by intent, with interested replies first. You can read and edit queued messages on the Actions page before they send. Kinlead sends from your own LinkedIn account through Unipile, and it never stores your LinkedIn password.
Lead search uses LinkedIn's standard search, so a Sales Navigator seat is not required. Kinlead keeps leads in lead groups, and any campaign can point at any group.
Use the API for reads and drafts first. Add write actions only after you have a cursor and a log of who approved each message.
Frequently asked questions
Related
Get started with Kinlead
Prefer markdown? How to monitor LinkedIn replies via API.md