Summary

Two tools, one job, media included.

This does not try to be a LinkedIn client — there is no feed reading and no analytics. It publishes, and it handles the part most small scripts skip: uploading an image or a video and attaching it to the post rather than dropping in a link. The setup cost is a LinkedIn Developer app, which is a one-time step but not a two-minute one.

What it is

A small LinkedIn publishing server. It runs the OAuth2 sign-in once, stores the token, and then posts on your behalf.

What you get

`authenticate` runs the LinkedIn OAuth2 flow and stores the resulting token
`create_post` writes a post and, given a local file path, attaches the image or video to it
Post visibility is a parameter — public, or connections only

Requirements

The PyPI package linkedin-mcp, version 0.1.7, installed with pipx and launched as `linkedin-mcp`. You supply a LinkedIn Developer app with the "Log In to LinkedIn" and "Share on LinkedIn" products enabled, then set LINKEDIN_CLIENT_ID, LINKEDIN_CLIENT_SECRET and LINKEDIN_REDIRECT_URI — the redirect URI is registered as http://localhost:3000/callback.

Setup effort

One command plus a key — pipx install linkedin-mcp, then supply credentials