repro-lens
Static reproducibility checks, experiment replay and before/after evidence for ML code and coding agents — no ML dependencies for scans.
Details
- External ID
- 1368134140
- Source
- GITHUB
- Company
- —
- Product
- repro-lens
- Website domain
- github.io
- Launched
- Sept. 13, 2026
- Cohort
- —
- Upvotes
- 9
- Upvotes percentile
- 0.21822956699974377
- Tags
- agent-skills, code-analysis, coding-agents, data-science, lightgbm, machine-learning, ml-testing, mlops, pre-commit, python, pytorch, pytorch-lightning, randomness, reproducibility, reproducible-ml, reproducible-research, scikit-learn, static-analysis, tensorflow, xgboost
- Fetched at
- Sept. 17, 2026, 5:03 p.m.
- Updated at
- Sept. 17, 2026, 5:03 p.m.
Enrichment
- Theme
- AI coding agents and developer tools
- Vertical
- Horizontal
- Function
- Observability & eval
- Audience
- Developer
- AI stance
- AI feature
- Project type
- Commercial product
- Normalized one-liner
- reproducibility checker and experiment replay tool for ml code
- Manually corrected
- False
Could you build this?
Partial While AST-based static analysis for seed detection is manageable, building a comprehensive, deterministic ML replay and reproducibility verification tool across major frameworks is technically involved.
What it would actually take: The engine requires a custom Python AST/bytecode parser that inspects framework-specific APIs (PyTorch, TensorFlow, scikit-learn) for non-deterministic operations (unseeded RNGs, nondeterministic CUDA operations, unpinned data loaders). Building a robust replay system also requires intercepting system-level RNG states, library versions, and floating-point tolerances across hardware variations.
Competitors
Other products that read as similar to this one — 1365 launches clear the similarity bar, closest 8 shown.
Attention rank: #997 of 1366 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 319 days after the earliest competitor.
- sloppy · github · 2026-09-10 · 13 upvotes · similarity 0.54
- CLI tool for detecting non-exact code duplication with embedding models · hn · 2026-07-02 · 91 upvotes · similarity 0.54
- Vho · hn · 2026-01-05 · 7 upvotes · similarity 0.54
- VibeDrift · hn · 2026-06-25 · 5 upvotes · similarity 0.54
- WideSWE · github · 2026-09-27 · 16 upvotes · similarity 0.53
- Mindwalk · hn · 2026-07-12 · 162 upvotes · similarity 0.53
- agenteval · hn · 2026-04-03 · 7 upvotes · similarity 0.52
- sediment · github · 2026-09-21 · 8 upvotes · similarity 0.52
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a observability & eval tool for Media & entertainment yet.