Where should I start?
Read the pccx docs for architecture context, then choose a focused repository issue. Documentation fixes, small tests, and reproducible bug reports are the lowest-friction first contributions.
Static preview
Assistant preview
This page shows the shape of a future project assistant using local, scripted content. It does not call a live model, does not use a provider endpoint, and does not include a Cloudflare Function.
Suggested questions
Which repository should I use?
Use pccx for documentation, the FPGA repository
for RTL and hardware evidence, pccx-lab for
CLI/core verification workflows,
pccx-llm-launcher for launcher scaffolding, and
systemverilog-ide for editor workflows.
Can I claim hardware results?
Only with reviewed, reproducible evidence. Claims about board inference, bitstreams, timing, or throughput need logs, reports, commands, commit identifiers, and the exact hardware and model setup.
What is required before a live assistant?
A live assistant needs a server-side boundary with Turnstile, rate limiting, Cloudflare AI Gateway, request validation, provider budget guards, and daily or monthly spend limits. None of that is active on this static preview.
Boundary
No live assistant is deployed here.
The current site contains HTML and CSS only. There is no secret, no environment variable, no provider SDK, and no route that can spend money.
Future work should add the assistant behind a reviewed Pages Functions boundary and a documented abuse-control plan before any provider traffic is enabled.