Numbers, made clearNumbers, made clear

Turns a plain-language business question into a runnable analytical SQL query built on your project's real schema, with an explanation.

Builds runnable cohort retention SQL on your real schema, plus a guide to reading the retention triangle and spotting misleading patterns.

Cleans, dedupes and reshapes a CSV or Excel file with a reusable Python script, saving a clean copy and leaving the original untouched.

Writes a Markdown dashboard spec covering audience, key questions, metrics, chart types and layout, ready to build in any BI tool.

Defines a YAML data contract between the producer and consumers of a dataset: schema, freshness and quality SLAs, and versioning rules.

Audits tables or pipeline output for nulls, duplicates, invalid values, orphaned keys and drift, with runnable SQL and a ranked findings report.

Generates dbt models in staging and marts layers with tests and column docs, following the conventions your dbt project already uses.

Scaffolds an ETL or ELT pipeline that is safe to re-run, fails loudly, and uses the data tools already in your repo.

Designs a YAML product analytics tracking plan with events, properties and naming rules, grounded in your app's actual code.

Builds conversion funnel SQL on your real event data, with drop-off per step and per segment, plus an insights template for next steps.

Turns a vague metric like "active users" or "churn" into a document with formula, grain, filters, edge-case rulings and a named owner.

Designs a star or snowflake warehouse schema with fact and dimension tables, dialect-correct DDL, an entity diagram and a design document.

Generates a Mermaid ER diagram from your database schema or ORM models, with columns, keys and relationship cardinality.

Generates migration files for a schema change, with both up and down versions and safety checks for risky table alterations.

Creates or updates your schema.prisma from your requirements or existing database, and gives the migration and client generation commands.

Analyzes a slow SQL or ORM query, spots missing indexes and N+1 problems, and suggests index and rewrite fixes.

Designs or reviews a database schema covering normalization, keys, constraints and indexes, and delivers SQL plus an ER diagram.

Generates realistic seed data for your database, with valid relationships and edge cases, in your ORM's format and safe to re-run.

Shows how to use z-ai-web-dev-sdk to read a URL and get its title, main HTML and publication time, from backend code or the CLI.