What is leaklane?

LeakLane is a tool for finding secrets, like passwords or API keys, that were accidentally committed into a group of code repositories, and it is built to run entirely on your own machine rather than sending your code to an outside service. Under the hood it uses an existing scanner called Gitleaks to search through repositories for exposed secrets, and it can optionally connect to a local AI tool called LM Studio to summarize the results, flag likely false alarms, and suggest what to fix first. Everything, including the AI step, stays local. The idea is to manage many repositories at once rather than scanning one at a time. You build what the project calls a pitlane, a saved list of GitHub and Bitbucket repositories, either by searching for them or adding their URLs directly. For private GitHub repositories, LeakLane reuses your existing GitHub CLI login instead of asking you to store a token inside the app. Every scan you run is saved, so you can look back at past results, compare what changed between scans, and browse a visual map of a repository's branches, tags, pull requests, and where the flagged findings sit within that history. It ships as a normal SvelteKit web app you can run locally, or as a desktop app built with Tauri, which packages a Python backend behind the same interface and can be built into a downloadable macOS package. To run it you need Git, Gitleaks itself, Python 3.10 or newer, Node.js and npm, and Rust if you want to build the desktop version. The GitHub CLI and LM Studio are both optional, used only for private repo access and AI summaries. Reports are saved as plain files, either in a local folder during development or in the app's data directory once installed. The project is currently an early alpha release focused on interface polish, and the author notes the macOS build is not yet signed or notarized. The README does not state a license.