The Shower Project That Just Made Every Database One Command

🚀 Quick Take

A new Python library called alchemy-utils hit alpha on August 12, 2026, and it could quietly change how developers — and traders who build their own tools — interact with databases. What started as a literal shower thought became a working prototype in a single morning, built by AI coding agents. The core idea: take the simplicity of sqlite-utils and make it work across PostgreSQL, DuckDB, and other engines through SQLAlchemy.

Here's the headline number: a one-liner that pulled every row from a table in a local PostgreSQL copy of a blog's database. No schema setup, no config files, no boilerplate. That's the entire pitch in one command.

🛠 What It Is

The project was sparked by Simon Willison, who wondered what a database-agnostic version of his popular sqlite-utils library might look like. sqlite-utils is beloved because it turns SQLite operations into dead-simple Python methods and CLI commands — insert, upsert, create, update, table introspection — without deep SQL knowledge.

The twist: instead of building it from scratch, Willison tasked two AI coding tools — Codex and GPT-5.6 Sol Ultra — with a research spike. The prompt was specific: build a library with the same core API as sqlite-utils, but backed by SQLAlchemy so it works with multiple database engines. The agents were told to use uv init for project setup, red/green TDD with pytest, and commit early and often. Credit the source via Simon Willison.

The result shipped as an alpha in remarkably few follow-up prompts. Two examples from the release show the flavor:

• List rows from PostgreSQL: uvx --with 'alchemy-utils[postgresql]' alchemy-utils rows 'postgresql+psycopg://simon@localhost:5432/simonwillisonblog' redirects_redirect
• Load a CSV of every tree in San Francisco into DuckDB, with schema auto-created: curl 'https://raw.githubusercontent.com/simonw/sf-tree-history/refs/heads/main/Street_Tree_List.csv' | uvx --with 'alchemy-utils[duckdb]' alchemy-utils insert 'duckdb:////tmp/trees.db' trees - --csv

That second one originally took nearly an hour to run; Codex optimized it down to around 35 seconds.

🧠 Why Traders Should Care

If you're purely a chart-watcher, an alpha-stage CLI tool might sound like developer noise. But the patterns here matter for anyone building crypto tooling, running backtests, or aggregating data across exchanges and chains.

First, this is a live demonstration that AI coding agents can now take an existing library's proven API design and port it to a new backend in hours, not weeks. The velocity of tooling is accelerating — what used to be a weekend project for a skilled developer is now a shower thought shipped before lunch. For traders who build their own scanners or dashboards, that means the tools you want will keep getting cheaper and faster to produce.

Second, the multi-engine angle is genuinely useful. Crypto data lives everywhere: PostgreSQL for order books, DuckDB for local analytics, SQLite for lightweight trackers. A single API that works across all of them removes friction when you're stitching together data sources. The auto-schema feature — where a CSV defines its own table structure — is especially handy for ingesting token metadata, on-chain exports, or exchange dumps without hand-writing schema migrations.

Third, the optimization story — an hour down to 35 seconds — is the kind of real-world gain that compounds when you're processing streams of token transactions or holder snapshots.

That said, this is an alpha. It's not battle-tested on production crypto workloads, and you shouldn't build your main infrastructure on pre-1.0 software. But it's worth watching — and if you're comfortable with Python, the one-liner examples are low-risk to experiment with on a throwaway database.

⚡ Put It To Work Today

If you want to try the tool yourself, grab a CSV you care about and run the DuckDB one-liner from the source. You'll have a queryable database in under a minute.

If you'd rather skip the coding entirely and put your time into actual research, here's the free alternative: the network's alerts arrive pre-screened with security data attached. The @gmgnalerts Telegram feed runs 450+ groups with live buy/sell alerts across SOL, BSC, ROBINHOOD and more. Each alert passes a layered security gate — GoPlus, RugCheck, GMGN entrapment/bundler/holder analysis, LP lock-burn checks — and the risks are printed directly on the alert as warnings. You see the red flags before you ever click a link.

For token tracking, XTRACK (@xtrack1bot) follows every alerted token and pings milestone movements with holders, LP status, and security data. @VBMBbot handles multibuy scanning. If you want the full picture in one place, blackhat.finance is a free web terminal with live trenches, trending, and a DYOR Academy library.

For actual trading execution, the network's alerts deep-link into GMGN — a fast memecoin terminal with sniping, wallet tracking, and PnL tools. You can register free at gmgn.ai/?ref=10Xboost, and it works with the GMGN Android app.

🎯 Bottom Line

alchemy-utils won't change your trading day tomorrow. But it's a signal: the gap between "idea" and "working tool" is collapsing. AI coding agents built this in hours, and the result is a genuinely useful utility that simplifies multi-database work.

The trader's takeaway is practical: you don't need to build your own data infrastructure from scratch, and you don't need to trust unvetted tokens blindly. Free tools exist right now for both sides of the job — pre-screened alerts and security-wrapped tracking for your research, and modern terminal tooling for execution. The research layer is already handled; the question is whether you're putting it to work.

—

Blackhat Empire — free multi-chain crypto alerts: @gmgnalerts XTRACK milestone tracker: @xtrack1bot | Multibuy scanner: @VBMBbot Web terminal + DYOR Academy: blackhat.finance GMGN terminal (free signup): gmgn.ai/?ref=10Xboost Not financial advice. Always DYOR.

🏴 Blackhat Empire — Free Alert Network

🚪 Telegram Portal: @gmgnalerts 📲 Trade on GMGN: gmgn.ai 📍 Live plays & full DYOR: blackhat.finance 🏴 Add all 7 MAIN groups: t.me/addlist 💬 Community Chat: @gmgnx_chat 🤖 Power tools: @VBMBbot · @xtrack1bot