act101

What is act101?

act101 is a deterministic merge gate for AI agent-generated code. As AI agents write code faster than teams can maintain it, architectural drift that once took five years now happens in five weeks. act101 closes that loop: your agent does the work, then the act gate classifies the change, finds the tests that reach it, diffs the side-effects, and returns a deterministic MERGE / REVIEW / BLOCK verdict with a signed receipt and one-command rollback.

Key Features

• Analyze — 41 analyzers detect dead code, cycles, coupling, hotspots, and seams before you touch a line.

• Act — 183 refactors (rename, extract, inline, move, modernize) across 163 grammars, all in one native Rust binary.

• Attest — 6 verify ops including classification, test reachability, side-effect diff, verdict, and signed receipt.

• Deterministic merge gate — Never rubber-stamp an AI edit again. Every change gets a provable verdict.

• One-command rollback — act history undo undoes operations and restores files instantly.

• ~85% token savings — Answers in one call instead of reading whole files.

• MCP & CLI — Integrates natively with Claude Code, Cursor, Codex, Windsurf, Zed, and opencode.

• Zero dependencies — Single native binary, no supply chain risk, no data exfiltration.

Who is it for?

act101 is built for three personas: individual engineers who want to verify every refactor their agent makes before merging; codebase owners and architects who need to see structural drift and prove refactors on every commit; and teams just here to build who want their AI to code better without deep technical setup. It also offers act101 online, a GitHub-native health scanner that grades AI-authored repos on Security and Architecture, landing results as PR comments, Check Runs, and SARIF uploads.

Why act101 stands out

Unlike traditional linters or review tools, act101 is purpose-built for the AI age. It classifies each change as format-only, behavior-preserving, or behavior-altering — and only auto-merges safe changes. When behavior changes and tests cover it, the change is routed for human review. When behavior changes without tests, it blocks. The result: compounding leverage instead of compounding debt. Your codebase stays shippable at agent speed.

Pricing

Builder Edition ($0) — free forever for personal, non-commercial, and open-source use. Includes 15 query operations, 5 codebase analyzers, rename + fix-auto, and 138 grammars. Engineering Edition ($29/mo) adds extract, inline, wrap, introduce, generated constructors & accessors, refactor verification, and commercial license. Architecture Edition ($59/mo) adds structural refactors, code scaffolding, taint + git overlays, and team licensing. Elite Edition ($99/mo) adds 25 premium languages including COBOL, Fortran, Ada, Verilog, Solidity, Vyper, CUDA, GLSL, HLSL, and WGSL. All paid editions include a 7-day free trial.