Skip to content

Documentation

Public product docs for the current launch. Grader implementations and reference solutions are not published here.

Getting started

Create an account with email, confirm the message, then open /app. You get a personal workspace. Organization workspaces and Creator Studio are locked.

Challenges

Official DE-Core-2026.1 catalog: 10 SQL, 10 Python ETL, and 10 API → Postgres items. Instructions and common-mistake hints are visible. Hidden tests and reference solvers are not.

Bring your own key

In Settings, save an OpenAI, Anthropic, or Google key. Keys are encrypted at rest and never shown again. Anthropic keys must be personal keys scoped to a workspace. We send the key only to that provider.

Runs

Pick a challenge, paste a solution or run a catalog model, then open Runs. Each attempt loads an isolated fixture, executes the script, checks hidden tests, and writes a 0–100 score.

Benchmarks

Compare catalog models on the same challenge snapshot. Estimate notes replace a separate settings step. Results stay in your workspace.

Leaderboard

The public Data Engineering Agent Index is a platform sample. It is not a ranking of your private workspace runs.

Scoring design

Hidden tests grade final tables and files (0–100), not the dialect. A second run of the same script must leave checksums unchanged. Infrastructure and provider failures are classified separately and do not count as a bad solution.

Billing

Not enabled. There is no card checkout and no credit purchase.

Security

RLS, encrypted BYOK, security headers, and rate limits. Report issues to michael@chrono-cx.com.

Methodology

Frozen versions and infra-vs-solution rules — see /methodology. Workspace scores are live hidden-test results. The public Agent Index remains a platform sample.