jev-curate
High-throughput dataset filtering using TypeSafe Jev typed judgments
Visit jev-curate.vercel.app ↗What is jev-curate?
jev-curate is a Rust-based pipeline for filtering synthetic and pretraining datasets using TypeSafe Jev's typed evaluation system. It streams Parquet and JSONL files through judgment rubrics, keeps passing rows, and writes rejects to separate files. Built for developers curating training data at scale, it offers three presets (reasoning-math, anti-sycophancy, code-correctness) and costs approximately $4.20 per 100M input tokens. The tool includes safety features like token bucket rate limiting, dry-run verification, and zero runtime dependencies, running as a single Rust binary or Python package.
Momentum
Daily activity over the last 30 days: upvotes, unique visits, new GitHub stars and Hacker News points.
How do we know it's vibe coded?
vibetrend discovered jev-curate automatically. Evidence: Rising on GitHub: agent files (claude-code) in AkashPriyadarshii/jev-curate. View source.
Similar vibe-coded apps

Open-source platform for orchestrating data, AI, and infrastructure workflows

Benchmark suite for code agents controlling robots in physical tasks

Open-source, cookieless web analytics with session replay and Core Web Vitals

Blind-graded benchmark of AI coding models on real planted bugs


Remote MCP server connecting AI agents to data from 91 platforms

Graph database on object storage, built for AI agent memory and context

Self-hosted SEO platform with site audits, rank tracking, and AI answer