quel 0.1.0

Open-source data quality that just works. No rules to write.

Quel

Open-source data quality that just works. No rules to write. Understands your data from day one. Gets smarter over time.

pip install quel

This project is in early development. Star the repo to follow along.

---

The Problem

Most data teams build ETL pipelines with no data quality framework at all — not because they don't care, but because every existing approach demands too much upfront investment.

Rule-based tools like Great Expectations and Soda ask you to write expectations for every column on every table. With 200 tables and 20 columns each, that's 4,000 rules to write and maintain. SaaS observability tools like Monte Carlo and Anomalo are easier, but cost $50–100K+ per year and require your data to leave your environment.

So teams do nothing. Bad data reaches production.

Quel makes data quality automatic.

How It Works

Three layers of protection activate progressively — none require writing a single rule.

Layer 1 — Format & Business Rules (Day 1) Quel reads your schema and immediately infers what each column should look like: emails are valid emails, amounts are positive, IDs are unique, dates aren't in the future.

Layer 2 — Statistical Anomaly Detection (Week 2+) After profiling runs accumulate, Quel builds statistical baselines automatically and alerts when distributions shift beyond expected variance.

Layer 3 — Business Metric Monitoring (Month 1+) Tracks business-level aggregates across dimensions to catch problems that column-level checks miss.

License

Apache 2.0