Observed arrival · 2026-10-29
proof-core: results a skeptic can rerun
An open-source tool that seals computed results to their inputs, code, and settings so others can verify and reproduce them offline.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Developers, scientific computing teams, and V&V reviewers
- Worth noticing
- The page says a standard-library verification file imports nothing from the producer, with no account or installation required to check a bundle.
Field notes
The page frames a result as a bundle containing the number, inputs, method, and signature, then treats integrity checking and reproducibility as separate tests. Its examples name concrete inputs: an AI evaluation can be tied to its prompt, model, weights, and seed, while a regulated margin check includes the limit set before the run and the rule. The project links to its GitLab source and describes a Python quick start.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue