PyDataUI Launch Announcement: LinkedIn Post
Copy and paste the text below directly into your LinkedIn post.
Stop rewriting your Python data prototypes in React just to make them fast enough for production. 🛑
For years, Python engineers have been caught between two extremes: 1. Streamlit & Gradio: Great for a 1-day proof-of-concept, but their "full-script rerun" architecture chokes on real datasets, causes memory leaks, and lacks commercial UI flexibility. 2. React / Next.js + FastAPI: Fast and scalable, but now your data team is forced to maintain dual repositories, deal with TypeScript build chains, npm dependencies, and state-sync boilerplate.
Today, we are changing that.
Introducing PyDataUI — a production-ready, full-stack Python framework engineered for Data & AI applications, powered by a compiled Rust HTTP core. 🦀🐍
💡 Why is PyDataUI different?
⚡ Compiled Rust HTTP Engine (PyRustAPI):
Serves up to 112,000+ requests/sec with sub-millisecond P99 latency. That is up to 60x faster than traditional Python UI rerun frameworks, with a base memory footprint under 30MB.
🎯 Fine-Grained Reactive State:
No full-page script re-executions. When your @s.action updates state, PyDataUI calculates the delta and updates only the affected DOM element via lightweight HTMX swaps.
🎨 80+ Modern shadcn/ui Inspired Components: Enterprise-grade design tokens out of the box — KPI metric cards, data tables, Plotly charts, AI chatbots, sliders, dialogs, badges, and filters.
📊 Universal Data Sniffing:
Pass a DuckDB relation, Polars DataFrame, PyArrow Table, or Pandas DataFrame directly into DataTable(s.data). No serialization or schema boilerplate required.
🖥️ Dual Execution: Web Browser & Terminal UI (TUI):
The same Python code runs as a modern, responsive web application or as an interactive terminal dashboard over SSH jump boxes with zero code modifications (pdu.run_tui(app)).
🔒 Enterprise Security Native: Built-in Role-Based Access Control (RBAC), JWT session management, PBKDF2 password hashing, and machine-to-machine API key management.
💻 A Full Reactive Dashboard in 20 Lines of Python:
import pydataui as pdu
app = pdu.App(title="Cluster Telemetry", theme="slate")
s = pdu.state(active_nodes=12, latency_ms=8.4)
@s.action
def scale():
s.active_nodes += 2
s.latency_ms = round(s.latency_ms * 0.92, 2)
@app.page("/")
def home():
return pdu.Container(
pdu.Heading("Distributed Cluster Telemetry", level=1),
pdu.Row(
pdu.Card(pdu.Text("Active Nodes"), pdu.Heading(s.active_nodes, level=2)),
pdu.Card(pdu.Text("Latency"), pdu.Heading(f"{s.latency_ms} ms", level=2)),
pdu.Button("Scale Up", on_click=scale)
)
)
if __name__ == "__main__":
app.run(port=8000)
No JavaScript. No Webpack. Just blazing-fast, reactive Python.
🔗 Try It Out Today:
We are 100% open-source under the MIT license:
⭐ GitHub Repository: https://github.com/rajboopathiking/pydataui
📖 Full Documentation & Tutorials: https://rajboopathiking.github.io/pydataui/
📦 Get Started: pip install pydataui
What are your biggest pain points when deploying Python data applications to production? Would love to hear your thoughts in the comments! 👇