Launching today

Packet Highway | Network Investigator

Launching today

See the network. Follow the evidence. Decide what to check.

0 followers

See the network. Follow the evidence. Decide what to check.

0 followers

Watch a synthetic network change, then ask GPT-6 Astra what it may mean. Explore Observed facts, Inferred explanations and Unknowns. Select an interpretation to inspect its cited observations, then decide what to check next.

Packet Highway | Network Investigator gallery image

Payment Required

Launch tags:Artificial Intelligence•Data Visualization•Security

Launch Team / Built With

OpenAI

GPT-6 Astra Challenge

How did Astra change the scope or ambition of what you built?

Maker

Astra let me narrow the project and raise the ambition at the same time. I already had a broader PSEC network-visualization project with a lot of useful ideas, but rebuilding the whole system would have been the wrong move. Because Astra can reason over structured, incomplete evidence and return structured outputs, I extracted the strongest part: a visual network story where code owns the observed facts, Astra interprets what they may mean, and the human decides what to do next. That turned a broad security project into a focused, evidence-aware product I could actually ship and explain clearly.

Report

2h ago

Maker

📌

I built Packet Highway around a question: can AI make network behavior easier to understand while keeping facts and interpretation visibly separate? Start with a sample home network. Familiar devices establish normal activity. Then an unrecognized device appears. Code records what happened; GPT-6 Astra offers possible explanations and questions the evidence cannot settle. Click an interpretation to see the exact observations it cites and the relevant parts of the map. The final step is a suggested human check. The idea grew from an older PSEC concept. Security-review lessons led me to reduce it to its strongest part: a visible evidence boundary. I rebuilt the demo around Astra and used Claude Opus 5 as an independent development reviewer. Astra is the runtime intelligence; Opus was QA only. Everything here is synthetic. The app cannot see your network or act on it. Citation validation checks that an observation exists; deciding whether it supports a conclusion still requires judgment. Code establishes facts. Astra interprets them. You decide. What was clear when you tried it, and where did the distinction between an observation and an interpretation need more explanation?

Report

12h ago

OpenAI

GPT-6 Astra Challenge

How did Astra change the scope or ambition of what you built?

Maker

Report

2h ago

Maker

📌

Report

12h ago