◈ Product roadmap

Where we're going

Built in public. Updated weekly. Every milestone tied to a real engineering sprint. Click a phase to expand milestones.

NL → pipeline generationAI

Claude API + schema context + structured output + live DAG preview

PDL compiler v0

Lexer, parser, name resolution, basic type checker in Python

DuckDB execution runtime

Run pipelines locally with row count preview and structured error output

Auto-catalog from runs

Every pipeline run registers sources, sinks, and column lineage

AI expectation generationAI

Profile dataset → generate quality checks → inline expect blocks

NL catalog searchAI

pgvector semantic search across all catalog metadata

Pipeline health dashboard

Run status, duration trends, error feed — the observability landing page

Basic RBAC + secrets vault

User auth, workspace isolation, encrypted credential storage

This roadmap is updated weekly as we build. Follow us on GitHub to watch it happen in real time. Want to influence priorities? Join as a design partner.