Structured Labs

Join Structured Labs and Recce for Data Reboot on April 21 -- RSVP here: https://lu.ma/ro29lcyb!

We open-sourced a little utility we built for ourselves: Supawald — a minimal CMS for managing files in Supabase Storage. Drag & drop uploads, folders, auth, and a “publish” button that triggers site rebuilds. npx create-supawald my-app Code: https://lnkd.in/eTbPGxsk

Awesome example of a data app built with Preswald. Nice work!

Build interactive data apps with Preswald ✔️

🚀 Preswald OSS Launch Week Day 5: Support for static Fastplotlib Preswald now supports Fastplotlib, a high-performance, GPU-accelerated plotting library, as a rendering backend for static visualizations. Drop in your fastplotlib figure, and Preswald will display it directly in your app. These are WebGPU-rendered visuals built entirely in Python. - Visualize large datasets with GPU rendering - Scientific plots with ultra-low latency - Static 2D or 3D charts inside Preswald apps - Easily showcase results from Fastplotlib pipelines Many scientific teams are adopting Fastplotlib for performance. Now you can show those plots in your Preswald app, as static components, without exporting images or screenshots. from preswald import fastplotlib fastplotlib(fig) Huge thanks to Trevor Wilson, Bhavya Gada for this awesome community contribution! And thank you to Kushal Kolar from the fastplotlib team for your support. And that's a wrap to Launch Week! Try today: https://lnkd.in/eU7Dva7K

Really creative example of an app built with Preswald! Nice work!

🚀 Preswald OSS Launch Week Day 4: New AI Chat Widget You can now add an AI chat box to your app that talks directly to your data. Preswald now supports a new chat() component. A simple way to build RAG-style assistants over any data source defined in preswald.toml. How it works: - Pass in a source (CSV, Postgres, ClickHouse) - Preswald summarizes the data and sends it to OpenAI - A chat widget appears in your app with full context of that data Example: from preswald import chat chat(source="iris_csv") chat("pg_db", "earthquake_events") Now you can explore datasets by asking questions like: - “What are the top 5 rows by value?” - “Which columns have missing data?” Shoutout to Arnav Jaiswal for this fantastic community contribution.

🚀 Preswald OSS Launch Week Day 3: S3 CSV Data Sources One of the most common requests we’ve heard for Preswald: “Can I connect to a CSV file in S3… without downloading it first?” Today, that’s a resounding yes. Preswald now natively supports reading CSVs from S3-compatible storage — including AWS S3, MinIO, or even local emulators — with zero glue code. You configure the source in preswald.toml, and Preswald takes care of the rest: credentials, streaming, and loading into DuckDB — ready for querying and visualization. What this unlocks: - Apps on top of private data lakes - Instant dashboards powered by S3 buckets - Multi-source analysis across APIs, databases, and object storage Shoutout to Varnit Singh for this fantastic community contribution.

🚀 Preswald OSS Launch Week Day 2: Support for Matplotlib Preswald now supports matplotlib. You can now render matplotlib charts directly in your Preswald app using just Python. Whether you’re working with scientific plots, custom subplots, or domain-specific visualizations, Matplotlib now plays nicely inside Preswald’s reactive workflow engine. What this unlocks: - Use legacy Matplotlib code in modern data apps - Render static or complex visualizations with full control - Great for notebooks and scientific analysis ported to apps Matplotlib is still the go-to for many scientists, researchers, and engineers. Developers can incrementally migrate existing scripts into interactive dashboards. No need to learn a new charting library.. your plots just work Shoutout to Anirudh kohli for this fantastic community contribution.

🚀 Preswald OSS Launch Week Day 1: API data sources You can now connect to any REST API declaratively. The response is parsed and loaded into DuckDB, making it queryable with SQL in your app. What this unlocks: - Live dashboards from APIs (analytics, ops, weather, etc.) - Integration with internal microservices or third-party tools - SQL over API responses, visualized instantly in tables or charts ✅ Why it matters: - Eliminate brittle data-fetching code - Prototype tools on dynamic data quickly - Combine API + database + file data in one unified app Thanks to Lokesh Poluru Velayudham for the contribution! https://lnkd.in/eU7Dva7K