Available now

KLAUSSY

Run multiple repos. Mix any agents.
In one window.

For the 20× developer.

A desktop app for macOS, Windows, and Linux. Start sessions across one repo or many — same branch everywhere, any number of agents (Claude Code, Codex, Gemini, Copilot) in each, even different ones side by side — review PRs with AI, and get local tab-autocomplete that never leaves your machine.

Free to try. $99 lifetime access unlocks an access key for every platform.

Need a different platform? All downloads ↓

Not sure if your Mac is Apple Silicon or Intel? Click the Apple menu → About This Mac. If the Chip row says “Apple M1/M2/M3/M4” you have Apple Silicon. If it says “Intel”, pick the Intel build.

Sessions, your shape.

Stack several on one repo, fan one across many repos, or run a mix of agents inside a single session.

Three sessions — A, B, and C — each pointing to the same repo, api.
Many sessions, one repo. Run parallel sessions on the same repo — a feature here, a hotfix there.
One session fanning out to three repos: api, web, and infra.
One session, many repos. Pull every repo a change touches into one session — same branch everywhere, side by side.
One session fanning out to three agents: Claude Code, Codex, and Gemini.
One session, many agents. Mix any number of agents in a session — Claude Code, Codex, Gemini, Copilot — even different ones on the same repo.

What it does

Any agent, your choice

Run each session on Claude Code, OpenAI Codex, Google Gemini, or GitHub Copilot. Pick a default, switch per terminal, or run the same session in two agents side by side.

Sessions, your way

A session is whatever you need — one repo or many. Pick the repos it touches (even straight from your recent GitHub repos, cloned on demand); each gets the same branch and its own worktree, with any number of agents you choose, even different ones side by side. Run several sessions at once — including more than one on the same repo — and resume any of them later with one click.

Repo-aware agents

On session start, Klaussy maps your repo — its conventions, rules, and import graph (fan-in/out, cycles, endpoint chains) — and gives your agents that context to draw on. Plan, review, debug, and implement runs ground their work in how your codebase fits together, not generic advice. Refreshes as the repo changes.

Auto-debug CI failures

Klaussy connects to your PR's CI; one click pulls the failing job's logs and your agent returns root cause + suggested fix, applied straight to the worktree.

Full PR review surface

Pull in any PR, read the diff with inline comments, run an AI review that breaks into per-finding cards — ignore, implement, or append to PR.

Plan · Debug · Review

Every worktree has a dropdown that spawns a dedicated agent tab running Klaussy's guided Plan flow, a Debug pass, or a multi-phase PR review — same worktree, no context loss.

Inline AI — locally

Tab-autocomplete powered by qwen2.5-coder via Ollama. ~100ms latency. No code leaves your laptop.

Safe commits

Before code leaves your hands, Klaussy reviews the staged diff — leaked secrets, silent failures, debug leftovers, correctness landmines, plus your repo's own linter, on a per-lens scorecard. Wired to the Commit button and as git pre-commit/pre-push hooks, so it catches agent commits and pushes too. Fix first or commit anyway.

Built-in editor

Monaco editor with LSP diagnostics. Edit and commit straight from the diff panel. AI-generated commit messages optional.

See it in action

One session, every repo, your pick of agents. Here a single session spans two repos side by side — OpenAI Codex on the left, Claude Code on the right — while Klaussy builds a conventions-aware CLAUDE.md so both agents work from your codebase's real structure.

Klaussy running one session across two repos: an OpenAI Codex terminal on the left, a Claude Code terminal in the middle, and a Changes panel on the right showing a freshly generated CLAUDE.md with the repo's architecture, API routes, and tech stack.

AI PR review: each finding is its own card with severity, file:line location, the offending code, and a written critique. Per-finding actions: Ignore, Add to PR, Implement, or open a chat to Ask or Investigate — all run on your selected agent.

The Review tab listing 24 findings as cards; the top one is a Blocker (Correctness) on src/hooks/useSpeechRecognition.js with a code block, a detailed written explanation, and a row of action buttons: Copy, Investigate, Ask, Ignore, Add to PR, and Implement.

Debugging a failing CI check: click Debug on the red row and your agent pulls the job logs, pinpoints the cause, and proposes a fix. Ask follow-ups in the chat, then click Fix this to have the agent apply it in the PR worktree.

Klaussy's Checks tab on a failing 'Lint, format, test, build' run, with a Debug panel showing the agent's analysis and a concrete patch, a follow-up chat composer, and 'Fix this' and 'Open as task' buttons.

Review the full diff in-app: every changed file with add/remove counts, the unified diff, and inline comment threads you — or your agent — can reply to. The same surface you'd get on GitHub.

The Files tab of a pull request: the left rail lists 38 changed files with per-file plus/minus bars; the main pane shows the .github/workflows/ci.yml diff with an inline comment thread.

Any agent, plus a real editor. Run Claude, Codex, Gemini, or Copilot in the terminal with a built-in Monaco editor and LSP diagnostics right beside it — edit, then commit from the diff panel.

Klaussy with a Gemini CLI session on the left and the built-in code editor open on api.ts — with a minimap and inline diagnostics — on the right.

What you need

  • macOS 12 (Monterey) or later — Apple Silicon or Intel — Windows 10/11, or Ubuntu 22.04+ (other modern Linux distros generally work).
  • At least one agent CLI installed and authenticated — Claude Code, OpenAI Codex, Google Gemini, or GitHub Copilot. Klaussy orchestrates the agent you choose — it doesn't replace it.
  • GitHub CLI (gh) authenticated. Required for PR review features.
  • Ollama (optional). Only needed if you opt into local inline autocomplete. On macOS, Klaussy can install it via Homebrew; on Windows/Linux, install from ollama.com.

Lifetime access

One-time purchase. Unlocks an access key that works on macOS, Windows, and Linux — and includes every future update.

Team — Small

$349

5 seats. Lifetime.

  • 5 access keys for your team
  • All future updates included
  • One invoice for the whole team
Buy team license

Team — Large

$599

10 seats. Lifetime.

  • 10 access keys for your team
  • All future updates included
  • One invoice for the whole team
Buy team license

Klaussy is free to download and try. An access key is required to keep using it past the trial.

FAQ

Does my code get sent to third parties?

When you use an agent, prompts + repo context go to that agent's provider via the CLI you already trust — Anthropic (Claude), OpenAI (Codex), Google (Gemini), or GitHub (Copilot). GitHub operations go through your local gh. Inline autocomplete runs entirely locally via Ollama and qwen2.5-coder:1.5b — nothing per-keystroke leaves your machine. We don't run a server of our own.

Do I need a subscription for the agents?

You need whatever plan each agent CLI you use is configured for — Claude, Codex, Gemini, and Copilot all run on your own accounts. Klaussy doesn't bill separately for AI usage — your purchase is a one-time license for the app itself.

How does the access key work?

After purchase you'll receive a license key. Paste it into Klaussy → Settings → License once, on any of your machines. The key is tied to you, not to a single OS — the same key activates Klaussy on macOS, Windows, and Linux.

What does “lifetime” mean here?

One-time payment, no subscription. You get every future update to Klaussy at no extra cost for as long as the app exists. AI usage (your agents, GitHub) still runs on your own accounts.

Which platforms are supported?

macOS 12+ (Apple Silicon and Intel), Windows 10/11, and Ubuntu 22.04+ (other modern Linux distros generally work). All builds are signed where the platform supports it.

Why the 2 GB download prompt for inline autocomplete?

~500 MB is Ollama's runtime; ~1 GB is the qwen2.5-coder:1.5b model weights. You only see this prompt if you opt in — otherwise a free word-based completer handles Tab.

What happens to my data if I uninstall?

On macOS, remove ~/Library/Application Support/Klaussy and ~/Library/Logs/Klaussy. On Windows, remove %APPDATA%\Klaussy. On Linux, remove ~/.config/Klaussy and ~/.local/share/Klaussy. Ollama and its models persist independently — uninstall it through your package manager and remove ~/.ollama.

Is this open source?

The application itself is commercial, closed-source. Feedback and bug reports live in a public repo (this one). The bundled open-source components are listed in About → Licenses.

How do I report a bug or request a feature?

Open an issue in this repo, or drop it in the Discord. Both get read.

Get Klaussy.

macOS, Windows, Linux. Free to try — $99 lifetime access unlocks an access key for every platform.

Download Klaussy macOS · Windows · Linux