Auto-Claude

Executive Summary

Status note (June 2026): Auto-Claude rebranded to Aperant on March 7, 2026 and is pivoting from a community open-source framework to a commercial "AI coding platform" from Norwegian company Mikalsen AI AS, with paid tiers listed as "pricing coming soon." The public GitHub repo (14.3k stars) has had no pushes since March 23, 2026 — roughly two and a half months of public silence as of this update.

Auto-Claude is an open-source autonomous multi-agent coding framework that turns Claude Code into a "virtual dev team." Developers describe a goal, and agents handle planning, implementation, and QA validation through a Kanban-style interface, with up to 12 parallel agent terminals and automatic conflict resolution. The rebrand to Aperant repositions it as "the world's first AI coding platform that builds, ships, and maintains your product" — but whether development continues in private or has stalled is not publicly verifiable.

---

Product Overview

Auto-Claude takes the "autonomous" concept seriously. Rather than requiring developers to prompt and approve each step, it implements a full autonomous loop: describe your goal, and agents plan the work, implement it across isolated worktrees, run QA validation, and even handle merge conflicts automatically.

The Kanban interface provides visibility into agent progress across planning, implementation, and review stages. Each task moves through the board as agents work, with the ability to run up to 12 agent terminals in parallel.

The Aperant rebrand broadens the pitch beyond coding to full lifecycle coverage — planning, architecture, coding, testing, deployment, and maintenance — with links to GitHub and Discord describing an "open-source foundation" beneath the commercial platform.

Key Capabilities

Product Surfaces

---

Technical Architecture

Auto-Claude is built with an Electron frontend and Python backend for agent orchestration. It communicates with Claude Code CLI to run agents, using git worktrees for isolation. The security model includes OS sandboxing, filesystem restrictions, and a dynamic command allowlist. A pre-rebrand commit shows work migrating the backend from the Python Claude Agent SDK to the TypeScript Vercel AI SDK, suggesting a multi-model architecture shift for the Aperant platform.

Project Structure

Aperant/
├── apps/
│   ├── backend/     # Python agents, specs, QA pipeline
│   └── frontend/    # Electron desktop application
├── guides/          # Documentation
├── tests/           # Test suite
└── scripts/         # Build utilities

Key Technical Details

---

Strengths

• Truly autonomous — Agents work for hours without intervention; built for overnight/background execution
• Cross-platform — Full native apps for Windows, macOS, and Linux (multiple formats)
• Open source — AGPL-3.0 license; community can inspect, modify, and contribute
• Built-in QA — Self-validating quality loop catches issues before human review
• Smart merging — AI-powered conflict resolution reduces manual merge work
• GitHub/GitLab integration — Import issues, investigate with AI, create merge requests
• Linear integration — Sync tasks with Linear for team progress tracking
• Real adoption — 14.3k stars and an active Discord made it the most-watched open-source autonomous coding tool

---

Cautions

• Public repo gone quiet — No pushes since March 23, 2026; one HN commenter in May 2026 already flagged the stale update cadence
• Commercial pivot underway — The Aperant rebrand introduces Pro/Team/Enterprise tiers with undisclosed pricing; the free open-source experience may change
• Authentication policy risk — Anthropic's February 2026 policy prohibits using Claude Pro/Max OAuth tokens outside Claude Code, a concern raised against the project's subscription-based model in issue #1871
• Claude dependency — Originally requires a Claude Pro/Max subscription; tied to a single provider (the Vercel AI SDK migration may change this)
• Autonomous risk — Running agents for hours unsupervised can produce unexpected results
• Git-only design — Built around git worktrees; teams on alternatives like jj report friction
• AGPL license — Copyleft requirements may not suit all commercial use cases

---

What Developers Say

Community discussion is moderate but real — mostly on Hacker News; Reddit discussion was thin as of June 2026.

• "Auto Claude is a nice opensource repo to let Claude generate entire application from just one prompt... It actually created a working thing." — an HN commenter who built a hotel site over 4 days with 4 Claude accounts
• "[Aperant] did this during the peak of the macmini auto claude phase but last update was 5 months ago" — an HN commenter, May 2026
• "We are using jj with great success but tools like auto claude are designed around git and git worktrees. It's a shame" — an HN commenter

---

Pricing & Licensing

Licensing model: AGPL-3.0 open-source foundation; commercial Aperant platform pricing not announced as of June 2026

Hidden costs: Claude Pro/Max subscription ($20-100/mo) historically required — though Anthropic's OAuth policy now clouds that model

---

Competitive Positioning

Direct Competitors

When to Choose Auto-Claude Over Alternatives

• Choose Auto-Claude/Aperant when: You want agents working autonomously for extended periods and accept transition-period uncertainty
• Choose Intent when: You prefer structured spec-driven workflows with human checkpoints
• Choose Agentastic when: You want manual control over parallel terminal agents

---

Ideal Customer Profile

Best fit:

• Developers comfortable with autonomous AI execution
• Open source enthusiasts willing to run the existing AGPL release as-is
• Users on any platform (Windows, macOS, Linux)

Poor fit:

• Developers wanting fine-grained control over each agent step
• Teams that need an actively maintained tool with predictable release cadence
• Teams concerned about autonomous code changes
• Those needing clear commercial pricing today (Aperant pricing unannounced)

---

Viability Assessment

The project sits in an awkward transition: the open-source repo that earned 14.3k stars is publicly frozen, while the Aperant website promises a commercial platform with no launch date or pricing. Anthropic's OAuth policy change in February 2026 removed the economics (Claude Max subscription reuse) that made the original tool attractive, which likely forced both the rebrand and the architecture migration.

---

Bottom Line

Auto-Claude was the most ambitious open-source autonomous coding tool, and its Kanban-plus-worktrees pipeline remains impressive and freely available under AGPL-3.0. But the project is mid-pivot: rebranded to Aperant, repositioned as a commercial platform with unannounced pricing, and publicly inactive since late March 2026. Adopting it today means betting on a transition whose outcome isn't yet visible.

Recommended for: Tinkerers comfortable running the existing open-source release as-is, and developers who want to study a complete autonomous-agent architecture

Not recommended for: Teams needing an actively maintained tool, anyone requiring clear pricing, or users dependent on Claude subscription-based auth given Anthropic's policy change

Outlook: Uncertain — the 14.3k-star community is a real asset, but until Aperant ships with public pricing and the repo resumes activity, treat this as a project in suspended animation

---

Research by Ry Walker Research • methodology