LinearB MCP server: marketed, not documented

What is verifiable from outside the product and what is not, with fewer numbers than usual because LinearB's marketing pages returned HTTP 429 to every automated request.

Last verified 25 August 2026 · from The LinearB sitemap, its documentation search index, the MCP registry, npm and GitHub

This page is one of 90 in a directory of monitoring and developer tools. Each page checks what an AI coding agent can get out of the tool through its MCP server, the connector that lets the agent query the tool directly.

What is the LinearB MCP server?

LinearB is an engineering-intelligence platform. It reads from Git providers, issue trackers and CI to report cycle time, throughput, review latency and AI tool usage.

An MCP server over that would let an agent ask how long reviews took last sprint without opening a dashboard. LinearB's product pages point that way. They could not be read for this guide, so this stays a description of the product.

That is as far as external verification goes. The rest of this page records what could and could not be established, because an invented tool list would be worse than no tool list.

Where an MCP server usually leaves traces, and what LinearB leaves:

A hosted product feature legitimately leaves no package and no repository. Leaving no documentation page is the part that stands out, because that is where a customer would look.

The mapping below follows the product, not the connector, because the connector's surface is unpublished.

In practice, every claim in the table above is about the platform rather than the connector. No tool has been observed, so nothing here should be planned around without checking inside the product first.

What can an AI agent do with LinearB?

Nothing is published, so nothing can be listed here. LinearB markets an MCP server on its own domain and calls it a platform feature reached inside the product, with no tool names, no counts and no documentation page. The product pages imply engineering metrics: cycle time, throughput and review latency. That is an inference from marketing copy, not a tool list.

Can LinearB watch your own AI agents?

Nothing can be confirmed here, and the reason is the one that runs through this page: LinearB publishes no tool list, so no signal can be checked either way. Searched on 25 August 2026.

LinearB MCP server limits: where answers come back incomplete

LinearB publishes no tool list. That single fact produces the limits below.

How to configure LinearB MCP for agents

1. 1

   Ask LinearB directly for the tool list

   It is not published, and no external source has it.

2. 2

   Ask which metrics are exposed

   Reporting a metric in a dashboard and exposing it as a tool are different things.

3. 3

   Ask whether anything writes

   Nothing external establishes whether the connector only reads.

4. 4

   Ask how it authenticates

   No documentation describes a credential, a scope or an endpoint.

5. 5

   Treat third-party servers with care

   The only one found has no stars, no description and no push since 14 August 2025.

Paste this into your agent’s instructions

This is the one guide on this site with no observed tool list. The block exists so an agent does not present platform capability as connector capability.

When you answer using LinearB MCP tools, state:
- The exact tool you called and the metric it returned.
  LinearB publishes no tool list, so the name is the only
  record of what ran.
- The date range and the teams or repositories the metric
  covers.
- That engineering metrics describe a period, so a figure
  without its window means nothing.
Never state that data does not exist. State the tool you
called and what it returned.
Write your answer in ASD-STE100 Simplified Technical
English. Use short sentences with one idea in each.

Do you need the MCP server at all?

Not on the published evidence. LinearB's dashboards are the documented route to these metrics, and they are faster than describing a query.

Faros AI and Sleuth both publish material about their own connectors, and cover overlapping engineering-intelligence ground.

DORA metrics are reachable from the source systems directly. GitHub and GitLab both expose pull request and pipeline data through documented servers.

For deployment frequency specifically, the CI vendor usually answers it more directly than an aggregation layer does.

Ask LinearB for the tool list before planning around it. Nothing public will tell you.

LinearB MCP server setup

There is nothing to install, and that is the finding rather than a gap in this page. LinearB markets an MCP server on its own domain and publishes no way to obtain one.

Five searches on 25 August 2026 found no package, no registry entry, no repository and no documentation page. Both @linearb/mcp-server and linearb-mcp return 404 from npm.

LinearB describes it as a platform feature reached inside the product, so the route below is its own interface rather than a client configuration.

LinearB product  ->  MCP feature   (in-product, undocumented)

npm view @linearb/mcp-server   # 404
npm view linearb-mcp           # 404

Every other client

Each block below is the configuration for one client, with the file path and the key that client expects.

curl -sI https://linearb.io/platform/mcp-server | head -1   # 429

Read-only access, permissions and security

• The credential model is undocumented

  No external source describes what the connector authenticates with or what scope it carries.

• LinearB already holds broad access

  The platform connects to Git providers, issue trackers and CI, so anything exposing it inherits a wide view of engineering activity.

• Engineering metrics describe people

  Cycle time and review latency aggregate individual behavior, which makes access control a people question as well as a data one.

• Whether anything writes is unknown

  Nothing external establishes that the connector is read-only, so treat that as unconfirmed rather than assumed.

Troubleshooting

You cannot find setup instructions

They are not on docs.linearb.io. Its search index covers 442 pages and none mentions MCP. The marketing pages under linearb.io are the only public material.

The marketing page will not load in a script

It returns HTTP 429 behind a Vercel Security Checkpoint to automated requests. A normal browser session is the only way to read it.

You cannot install it

There is nothing to install. No npm package exists under @linearb/mcp-server or linearb-mcp, and there is no MCP registry entry.

A third-party server behaves oddly

The only one found is Sudhir-Dharan/LinearB-MCP-Server: MIT, zero stars, no description, and no push since 14 August 2025.

LinearB MCP server: Reference

What engineers report

There is no public engineering discussion of this connector, and no published inventory. What follows is the search that established that, which is the honest form of this section when the record is empty.

Five searches were run on 25 August 2026. The LinearB sitemap was parsed for MCP URLs and last-modified dates. The docs.linearb.io search index, covering 442 pages, was searched for MCP with no match. The MCP registry was queried for linearb with no result. Both @linearb/mcp-server and linearb-mcp returned 404 from npm. The GitHub API was searched for repositories matching linearb and MCP, returning one third-party project. The marketing pages themselves returned HTTP 429 to every automated request and were not read.

Should you connect an agent to LinearB?

Yes, though there is no packaged release. Best for a conversation with LinearB, rather than a plan built on published facts.

Strongest fit

Existing LinearB customers who can see the feature inside the product and confirm its tools with their account team.

Main advantage

The pages are current, so this is a live feature rather than an abandoned one.

Main weakness

Nothing about it is documented, registered or packaged, so no part of it can be checked before you commit.

Operational risk

Unknown. No credential model, scope or write capability is described anywhere public.

Frequently asked questions