3xcode is a command-line tool that converts legacy SQL (stored procedures, functions, views, and triggers) written in T-SQL, PL/SQL, PL/pgSQL, MySQL, Spark SQL, or generic ANSI SQL into production-ready PySpark. It runs a multi-phase AI pipeline with built-in validation rather than a single one-shot prompt, so the output comes with a record of what was checked, not just a promise that it works.
Who It's For
Data engineering teams
Teams and platform owners moving warehouse workloads onto Spark or Databricks, converting a real, dependency-heavy codebase rather than a handful of standalone scripts.
Platform owners
Whoever owns the migration timeline and needs visibility into progress, cost, and quality across a large batch, not just a working script on one developer's laptop.
Consultancies
Teams running repeated migrations for different clients, who need a repeatable, auditable process more than a one-off script.
How It Works
Point it at your code
A single file, or a whole directory of T-SQL, PL/SQL, PL/pgSQL, MySQL, Spark SQL, or ANSI SQL.
Multi-phase AI conversion
Discovery, planning, conversion, and validation run as distinct phases, in dependency order, not one black-box prompt.
Reviewed, runnable output
Production-ready PySpark plus a per-file quality report, graded and banded so you know exactly what to check before you ship it.
# Convert a single file 3xcode pyspark convert your_file.sql # Convert a whole directory as a batch, parallel workers auto-tuned 3xcode pyspark convert --dir ./sql/ --session "my-migration"
Prefer a visual dashboard? Run 3xcode ui
Everything the CLI does is also available in a local, browser-based dashboard: a guided batch wizard to start a run, live cards to watch it finish, and the same workspaces, sessions, and quality reports. Both drive the same underlying data, so use whichever fits the moment. Explore the Web UI.
See How It Works for the full 5-phase pipeline, and Sessions & Batches for how a migration is organized end to end.
Why Trust It
Migrating production SQL isn't a task to hand to a black box. Three things about how 3xcode is built are worth understanding before you point it at real code.
Where your code goes
The CLI runs on your machine. Your SQL is sent only to the AI provider you configure (Anthropic, AWS Bedrock, Google Vertex AI, or Azure AI Foundry, your own key with BYOK or the platform-provided one), directly from your machine to theirs. It is never sent to 3XDE's own servers: the 3XDE backend only handles authentication, licensing, and usage tracking, and never sees your SQL source or your generated PySpark.
Success-only billing
You're only charged for conversions that complete successfully. A file that fails, times out, or is never attempted because a run was cancelled doesn't count against your monthly allowance.
Enterprise controls
Organizations, an optional shared license and quota pool across your team, and account-level usage visibility, so a migration run by five people is still one thing an admin can see and manage.
Bring your own key
Route conversions through your own provider account instead of the platform-provided key. Your credentials are stored locally via your OS keychain, never on 3XDE's servers.
Every file gets a report
A confidence and completion grade, banded from Perfect down to Low, so you know exactly which files to review before trusting them in production.
See How It Works for the exact data flow, and Billing & Quota for how success-only billing and quota holds work in detail.
Get Started
Install and convert your first file
Install with pip install 3xcode-cli, log in, and run your first conversion in under 5 minutes.
Explore the Docs
Batch Conversion & Sessions
Convert a whole codebase in parallel, with adaptive workers, resumable sessions, and a per-batch quality report.
Scheduled Conversions
Run a batch later, on a cadence, or on infrastructure that needs to survive a reboot.
Web UI
A local, browser-based dashboard for the same batches, sessions, and settings the CLI manages.
Or jump straight to the full feature guide index.