Gitter. Two workflows are failing, and three pull requests are waiting on a review.

Gitter is a desktop dashboard for GitHub and GitLab. It reads the repositories you choose, tells you what needs you, and keeps all of it on your computer.

Example data

Overview

Needs attention

Repositories

    The Overview screen, with invented data. The sentence above it is built from the list the same way the app builds it.

    Each screen answers one question.

    Choose the repositories you care about and Gitter sorts what comes back into screens. Workspaces narrow every screen to one owner or organization.

    Who is waiting on me?

    Pull requests and Issues

    Review state, checks, reviewers and drafts for every open pull request. Filter issues by repository, label or milestone, and open a new one from the app.

    • Add retry budget to the sync clientApproved
    • Move session storage to IndexedDBChanges requested
    • Bump the GraphQL clientAwaiting review
    • Rewrite the log viewerDraft

    What is broken?

    CI/CD

    Workflow runs and pipelines with their duration, branch and commit. Open one to see its jobs and the step that failed, without leaving Gitter.

    Build and test main, 4m 12s Failing step: Run integration tests

    Did it ship?

    Deployments

    The latest state of each environment, plus its history. A repository that reports no deployments is shown as unavailable, never as failed.

    • acme/payments-api, productionSucceeded 2h ago
    • acme/payments-api, stagingFailed 20m ago

    acme/docs: unavailable. This repository reports no deployments.

    What happened while I was away?

    Activity

    Commits, pull requests, issues, runs, deployments and releases in one timeline. Gitter keeps 120 days of it on your machine.

    • Run #4182 failed on main14m
    • Deployment to production succeeded2h
    • Pull request #214 merged3h
    • 3 commits pushed to main5h
    • Release v2.4.0 published1d

    What am I doing next?

    Kanban

    Every repository gets its own board, stored on your computer. Custom columns, drag and drop, labels, priority and due dates, with an optional link to an issue or pull request.

    To do

    Rate limit banner copyPriority high, due Friday
    Linux keyring check

    In progress

    Log viewer searchLinked to pull request #231

    Done

    Run tick tooltips

    There is no Gitter server. Your computer talks to GitHub and GitLab directly.

    Gitter has no account of its own and no cloud database. The window you see never gets your token. A separate process on your machine makes every request.

    Your computer
    The windowShows your data. Asks for it, but never holds the token.
    asks for data
    Gitter's Go processMakes every API call and keeps the token in memory.
    OS credential storeWindows Credential Manager, macOS Keychain or the Linux Secret Service.
    API calls with your token
    GitHub and GitLabThe only servers Gitter talks to.
    A Gitter serverThere isn't one.
    The token stays out of sight
    It lives in your operating system's credential store. Gitter never sends it to the window, writes it to a file or puts it in a log.
    It reads everything and writes two things
    Gitter posts a comment on a pull request and opens an issue, and only when you start either one. It never merges, closes, edits or deletes.
    It works offline
    Cached data stays available without a connection. Gitter syncs when it launches, every ten minutes by default, or when you ask. Rate limits pause syncing until they reset.
    Your boards live on this device
    Clear the cache and Gitter fetches it again. Uninstall Gitter or wipe its data and your Kanban boards are gone, because there is no backup and no sync.

    Get Gitter running in three steps.

    1. Download it

      Pick the build for your system. Each one is a single download.

      The builds are not code-signed, so Windows SmartScreen or macOS Gatekeeper may warn you the first time you open one. Checksums are on the release page.

      To build it yourself, run wails build with Go 1.25 or newer, Node 20 or newer and the Wails v2 CLI.

    2. Connect an account

      Choose Connect GitHub. If you are already signed in with the GitHub CLI, Gitter uses that account. If not, paste a personal access token with the repo and read:user scopes. GitHub has no narrower scope for reading private repositories.

      GitLab connects with a personal access token.

    3. Pick your repositories

      Nothing is connected automatically. Choose repositories on the Integration page and Gitter runs its first sync. After that it only refreshes what has gone stale.

    No AI features, billing, cloud sync, webhooks or accounts.

    Gitter reads from the services you already use and keeps what it learns on your computer. That is the whole product.