Windows macro recording, screen-aware playback, and visual automation workflows.
Clawmation is a local Windows 10 and 11 macro recorder and macro player for games, including Roblox. It records mouse and keyboard input, replays editable macros, detects screen changes with image matching, colour detection, pixel detection, and OCR, and connects those actions through visual Loops. Download the current installer above or review the source, architecture, and portable file formats below.
- Macros: record mouse and keyboard input, edit metadata and repeat rules, manage presets, run at different speeds, and export individual recordings.
- Watch: detect a saved image, colour, or text on screen, restrict detection to a region, and click, press a key, or move the pointer when it appears.
- Loops: connect recorded macros, waits, screen checks, branches, recovery actions, and stop conditions in a reusable visual workflow.
- Safeguards: attach recovery guards and vision checkpoints directly to a macro without changing the original recording.
- Anti-AFK: briefly activate a selected game window, perform a configured action, and return focus to the previous application.
- Local operation: macro data, screenshots, OCR, and image recognition stay on the computer. Clawmation has no account system, telemetry, or cloud processing.
Loops turn independent recordings and screen checks into directed workflows. Each output is explicit, including If works and If fails paths, so error recovery and stop conditions remain visible.
The example above embeds a snapshot of the Daily Quest Route macro, waits for
a reward image, branches on the result, runs a recovery path when required, and
finishes through explicit stop nodes. Embedded macros remain independent from
later edits to the source recording.
The Macros workspace keeps recording, filtering, playback controls, repeat rules, presets, safeguards, notes, and run statistics in one view.
Watch can respond without starting a macro. A trigger combines one visual condition with one action, a confidence threshold, an optional screen region, and a cooldown. Testing shows the captured frame and the detected target before the trigger is armed.
Clawmation uses two versioned, compressed formats:
| Format | Contents | Use |
|---|---|---|
.clawmation |
One macro | Share or archive a recording |
.clawbundle |
A macro, safeguards, and referenced vision images | Move a complete screen-aware automation setup |
Both formats use a manifest with declared sizes and BLAKE3 digests. JSON payloads use Zstandard compression, duplicate images are stored once, and imports reject traversal paths, unsafe names, undeclared files, invalid digests, and excessive expanded sizes. Existing macros are never overwritten. See Portable file formats for the complete container and compatibility contract.
Recording, playback, screen capture, template matching, pixel detection, and OCR run locally. Captured content is not uploaded. The application contacts GitHub only to check the signed release manifest used by its updater.
- Windows 10 or Windows 11
- WebView2, included with Windows 11 and current Windows 10 installations
- A 64-bit Windows installation for the published installer
Development additionally requires Node.js 20 or newer, Rust stable with the MSVC toolchain, and the Visual Studio Build Tools workload Desktop development with C++.
- Download
clawmation_1.2.1_x64-setup.exe. - Run the installer.
- Launch Clawmation and configure the record, play, and emergency-stop shortcuts in Settings.
The application checks the signed Tauri update manifest at launch. Release assets and notes are available on the Releases page.
Install the frontend dependencies:
npm installRun the Tauri development application:
npm run tauri devBuild the frontend:
npm run buildBuild the Windows installers:
npm run tauri buildRelease bundles are written under src-tauri/target/release/bundle/. Debug
builds store local data in the repository working directory; release builds
store portable data beside the executable.
Run the frontend suite and TypeScript checks:
npm test
npx tsc --noEmitRun the Rust suite:
cargo test --manifest-path src-tauri/Cargo.tomlHardware tests that move the cursor, capture the live desktop, or display native overlays are ignored by default. Run those individually on a machine that is not in active use:
cargo test --manifest-path src-tauri/Cargo.toml --lib -- --ignored --test-threads=1| Path | Responsibility |
|---|---|
src/ |
React 19 and TypeScript interface, typed API seam, and view components |
src-tauri/src/commands/ |
Validated Tauri command surface |
src-tauri/src/engine/ |
Macro, Watch, guard, schedule, chain, and Loop execution |
src-tauri/src/hardware/ |
DXGI capture, Win32 input, recording hooks, OCR, pickers, and vision |
src-tauri/src/models/ |
Persistent configuration and file-format contracts |
src-tauri/src/shell/ |
Tray, global hotkeys, updater, overlays, and indicator windows |
Read Architecture for the dependency map and Design for the interface contract.
The vision layer is implemented in Rust without a Python or OpenCV runtime. It contains native implementations of the image operations used by the detector, including colour conversion, morphology, contours, resizing, CLAHE, Canny, and masked template correlation. Matching uses a fast local search around the last known location before broader multi-scale and edge-based passes.
Read CONTRIBUTING.md before opening a pull request. Changes to detection, input synthesis, persistence, portable formats, or updater behavior require focused regression coverage and a clear compatibility note.
Clawmation is source-available, not open source. Personal, non-commercial use of official binaries and private evaluation of the source are permitted under the Clawmation Source-Available License 1.0. Redistribution, commercial use, hosted use, rebranding, derivative distribution, and machine-learning training use are prohibited without prior written permission.
Versions previously released under the MIT License remain governed by the license that accompanied those versions.