Repository navigation
Draw a timed motion, fit a one-motor linkage, and inspect what happens between animation frames.
Sweepforge 0.1.0 ships an editable local studio, bounded crank-rocker fitting, full-cycle interval capsule/tool clearance across dimension tolerances, nominal collision witnesses and explicit unresolved results. Exports freshly recompute inputs and carry source, exact dependencies, an outside-directory verifier, offline playback and conceptual SVG/DXF bar patterns in millimetres.
Try Between 64 frames: the original synthetic tool-point example has +0.749 mm gap at every one of 64 nominal samples, yet the interval checker finds a collision between them. This diagnostic explicitly excludes bars. Sculpted loop checks all six bar capsules and the tool across ±0.05 mm dimensions.
Install the wheel with Python 3.12 or later, then run sweepforge studio and open http://127.0.0.1:6381/. From the source archive, use python -m pip install -r requirements.lock .. Calculations use no network; dependency installation needs a package source or local wheel cache.
Assets include the actual browser-exported diagnostic and the all-body design bundle. Extract either ZIP, install its requirements.lock in an isolated environment, and run python /absolute/path/to/bundle/verifier.py from another directory. offline.html shows the frozen nominal motion and saved evidence; playback does not recompute interval checks. sweepforge-SHA256SUMS.txt records the other asset digests.
Verified: 27 Python tests and three Node controls locally and on each of Linux/Python 3.12, Linux/Python 3.14 and macOS/Python 3.14; fresh installed CLI/inverse and portable replay on all three hosted jobs; actual browser path edits/fitting, draft preservation/refusals, downloaded replay, and mobile width checks. Independent geometry, interval, drawing and runtime controls are summarized in sweepforge-evidence.json. Reviewed implementation.
One fixed nominal 80 mm ground span, one planar topology and assembly branch, fixed circular obstacles. No hardware, self-collision, load/flex/joint-play or manufacturing safety validation. Flat patterns are conceptual. Fitting is finite and may be approximate; interval-library support is experimental. No novelty or superiority claim over established mechanism-synthesis systems. Development and independent review used Codex; the posted receipt is from the author account, not a separate human approval.