Skip to content
larasrinathPublic

About

A local-first workspace for LinkedIn candidate discovery, evidence review, and score-based comparison. Built on linkedin-mcp-server.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

Compass

Find candidates on LinkedIn, review the evidence behind their scores, and compare people against your role criteria. Searches and downloaded profiles stay on your computer, ready to revisit.

Get started

Download or clone this repository, open its folder in a terminal, and run:

./compass

On Windows, open the folder in PowerShell and run:

.\compass

Or double-click Start Compass.bat in the folder. It works from wherever the folder is, and keeps the window open if startup fails so you can read the error.

Compass installs what it needs, opens the app, and opens a LinkedIn sign-in window on first use. Sign in, complete Search criteria, then choose Run search. Later launches reuse your installation and login.

Requirements: macOS, desktop Linux, or 64-bit Windows 10/11; Git and internet access for setup. Windows uses its built-in PowerShell; no WSL, Python, or Node installation is needed. macOS/Linux also need curl. Linux needs Chromium's system libraries. The first launch may take a few minutes. Keep the terminal open; Ctrl+C stops Compass without deleting saved work. Running the same command again gracefully restarts the previous instance from this repository and opens the app. Closing the browser tab alone does not stop Compass.

Startup help · Developer setup

Compass candidate results with evidence summaries and review actions

Screenshots use fictional candidates in the working app.

From search criteria to comparison

  1. Set your search criteria. Edit the role description, skills, credentials, locations, and minimum experience on one page. Add optional preferences where useful, then choose Continue to search. The description is not automatically converted into criteria.
  2. Find candidates. Run a search. New profiles download and receive scores automatically with the default settings; existing downloads are reused.
  3. Review the list. Check names and duplicate sources, add a review note, then choose Confirm list & show ranking. Switch between Cards and Ranked list, with up to 30 profiles per page.
  4. Inspect and compare. Open Review to see a candidate’s score and saved evidence. In Compare matches, select two or three people and choose View comparison.
  5. Return to saved work. Saved searches keeps previous runs together. Open results returns to the full list.

Use Settings for download limits, pacing, and scoring weights. Changing weights recalculates saved evidence locally. The in-app How it works guide walks through the same workflow with interactive examples.

Downloads, scores, and your data

  • Batch size: up to 1,000 new profile downloads per search batch by default. This is not a limit on the total profiles you can save. LinkedIn may return fewer results.
  • Speed: one or two profiles download at a time, depending on Settings and connector support. A larger batch does not increase simultaneous downloads.
  • Reach: choose first-degree, second-degree, or third-degree and beyond in Search criteria. Leave all unchecked to search across networks. Connection distance does not affect scores.
  • Profile visits: retrieval opens LinkedIn pages in your signed-in session. Hidden browser tabs do not make visits anonymous.
  • Scores: scores reflect retrieved evidence against your criteria. Confidence reflects evidence availability. Neither verifies a qualification or predicts job performance; open the saved sources to assess the match.
  • Local storage: searches, profiles, and evidence are saved in ~/.linkedin-dashboard/session.db. You can review saved work while the connector is offline. Compass does not send messages or connection requests.

Screenshots

Search criteria, ranked results, candidate review, and settings

Search criteria

Search criteria with an editable role description, skills, locations, and minimum experience

Ranked results

Candidates ranked by score, with confidence and review actions

Candidate review

Candidate drawer with score and signal breakdown

Settings

Download controls and scoring weights

Saved searches

Saved search with a three-person preview

How it works

Interactive Compass guide

More information

Acknowledgments

Thank you to Daniel Sticker (@stickerdaniel), author of linkedin-mcp-server, and its contributors. Their open-source connector powers Compass’s LinkedIn access.

License

Compass is MIT licensed. The LinkedIn connector retains its Apache-2.0 license. Bundled fonts retain their SIL Open Font Licenses: Figtree and Plus Jakarta Sans.

About

A local-first workspace for LinkedIn candidate discovery, evidence review, and score-based comparison. Built on linkedin-mcp-server.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages