Aperant: A Desktop Multi-Agent Coding Tool That Is Pausing Pull Requests for a 3.0 Rewrite
Autonomous multi-session AI coding. If you modify and distribute it, or run it as a service, your code must also be open source under AGPL-3.0.
At a glance
- What is it?
- Aperant (formerly Auto Claude) is an Electron-based app that runs multiple Claude Code agents in parallel to plan, build, and validate software. The current 2.x line is in maintenance mode while the project rebuilds for 3.0, and pull requests are closed until then.
- Who is it for?
- Adopt Aperant 2.x if you already have a Claude Pro or Max subscription, work in a git repository, and want a visual kanban-style interface to run multiple coding agents in parallel without touching your main branch. Do not adopt it if you need new features soon, because the 2.x line only gets critical fixes and pull requests are closed.
- Can I use it commercially?
- Yes, with strict conditions. AGPL-3.0 is a network copyleft licence: if people use a modified version over a network, for example as a hosted service, you must offer them its source code under the same licence.
- Is it still maintained?
- Yes. The repository last received commits 94 days ago.
- What is it written in?
- Mainly TypeScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 14, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What Aperant Actually Does and Who It Serves
Aperant is a desktop application that automates multi-session AI coding. It targets developers who already pay for Claude Pro or Max and use Claude Code CLI. The README describes it as an autonomous multi-agent coding framework that plans, builds, and validates software. You describe a task, and agents handle planning, implementation, and validation. This is not a code editor plugin; it is a standalone Electron app that manages multiple agent terminals. The intended user is someone who wants to offload entire feature builds to AI agents while keeping their main git branch clean. It is not for teams that need a self-hosted CI replacement or a tool that works without a commercial Anthropic subscription.
The Mechanism: Git Worktrees, Parallel Agents, and a QA Loop
The key architectural choice is isolation. All changes happen in git worktrees, so the main branch stays safe while agents work. The app can run up to 12 agent terminals in parallel, each presumably operating in its own worktree. After agents finish, an AI-powered merge step resolves conflicts when integrating back to main. A self-validating QA loop catches issues before human review. There is also a memory layer that lets agents retain insights across sessions. This design means the tool depends heavily on git semantics; if your project does not use git, the tool likely cannot work. The README explicitly lists 'Git repository' as a requirement. The parallel execution is a real feature, not a marketing phrase, because it is tied to the worktree isolation model.
Getting It Running: Requirements and Quick Start
The README gives a clear five-step quick start. First, download and install the app for your platform. Second, open a git repository folder. Third, connect Claude via OAuth setup, which the app guides you through. Fourth, create a task describing what you want to build. Fifth, watch agents work. The hard requirements are a Claude Pro or Max subscription, Claude Code CLI installed globally with 'npm install -g @anthropic-ai/claude-code', and a git-initialized project. Downloads are available for Windows, macOS (both Apple Silicon and Intel), and Linux in AppImage, Debian, and Flatpak formats. All releases include SHA256 checksums and VirusTotal scan results. The beta download links point to version 2.8.0-beta.5, while the beta badge links to v2.8.0-beta.6, so you should verify which binary you actually get.
The Elephant in the Room: Maintenance Mode and the 3.0 Rewrite
The most important fact for any adopter is that the current 2.x app is in maintenance mode. The README states that the project is rebuilding the entire app from the ground up for Aperant 3.0, which adds cloud features on top of the local desktop experience. Development is happening in a separate repository, and code pull requests against the current app will be closed. This is a genuine limitation: if you find a bug in 2.x, you cannot submit a fix that will be merged. You can only report it via issues or discussions. The project says 3.0 will stay fully open source under AGPL-3.0, but there is no public release date. For a tool that automates coding, this pause in contributions is a significant risk. You are relying on the maintainer's internal timeline, which is not visible in the public repo.
Licence Implications: AGPL-3.0 and What It Means for Your Use
The repository is licensed under AGPL-3.0. The README is explicit: if you modify and distribute it, or run it as a service, your code must also be open source under AGPL-3.0. This is a strong copyleft licence. For an internal tool that you run locally and never distribute, the obligations may be limited, but if you build a service around Aperant or modify it and share the modified version, you must release your changes under AGPL-3.0. This is a real constraint for companies that want to embed the tool in a proprietary product. The project homepage is aperant.com, and the repo is on GitHub, but the licence is not a suggestion; it is a legal boundary. You should consult your own counsel for specific implications, but the README's warning is clear.
A Real Alternative: Running Claude Code Directly
The obvious alternative is to skip Aperant and use Claude Code CLI directly. The CLI is the underlying engine that Aperant orchestrates. With the CLI, you can run agents in your own terminals, manage sessions manually, and use your existing git workflow. The difference is that Aperant adds a visual kanban board, parallel execution with up to 12 terminals, isolated worktrees, and a memory layer. If you do not need those features, the CLI gives you the same AI coding capability without the AGPL-3.0 obligations or the maintenance-mode risk. The CLI also does not require you to adopt a separate app. For teams that already have a process for managing multiple coding sessions, the CLI is a lighter-weight path. Aperant's value is in the orchestration and UI, not in the AI itself.
What to Verify Before You Adopt
Before you invest time, verify three things. First, confirm that your Claude subscription tier is sufficient; the README requires Pro or Max, not the free tier. Second, test the OAuth connection with your Claude Code setup, because the app depends on that integration working smoothly. Third, check the beta download naming inconsistency: the badge points to v2.8.0-beta.6, but the download links are for 2.8.0-beta.5. If you download the beta, confirm you are getting the version you expect. Also, since the project is in maintenance mode, check the issues and discussions to see if any known bugs affect your workflow. The README says the 2.x app is stable and usable, but 'stable' in maintenance mode means no new features, not that every edge case is solved.
Editorial conclusion
Adopt Aperant 2.x if you already have a Claude Pro or Max subscription, work in a git repository, and want a visual kanban-style interface to run multiple coding agents in parallel without touching your main branch. Do not adopt it if you need new features soon, because the 2.x line only gets critical fixes and pull requests are closed. Before committing, verify that your project can accept AGPL-3.0, confirm that the beta release you choose matches your platform (the beta downloads are labeled 2.8.0-beta.5 while the badge points to beta.6), and check that the OAuth flow with Claude Code works in your environment. The 3.0 rewrite is the future, but its release date is not public, so treat 2.x as a stable but frozen tool.
Community notes