I’ve been building PRISM - a self-hosted OSINT toolkit you run yourself instead of pasting investigation targets into someone else’s web service.
Give it a domain, IP, email, phone, or username and it runs 22+ modules in parallel into one dashboard: WHOIS, DNS, crt.sh subdomains, GeoIP, threat intel (Shodan/VirusTotal/AbuseIPDB/Censys), breach data, username search across 3000+ sites (Blackbird + Maigret), dark-web mirror checks, and more. Results come with an entity graph, a GeoIP map, an OPSEC exposure score (0–100), and HTML/PDF/CSV/Markdown exports.
14 of the 22 modules work with zero API keys (missing keys degrade gracefully instead of erroring).
Stack: FastAPI + Next.js 14, runs with one docker compose up. MIT licensed.
Demo: https://getprism.su/ Github: https://github.com/NovaCode37/Prism-platform
Built it solo - feedback welcome, especially on which modules you’d want added.
Did the LLM choose the name? There’s an obvious existing semantic link between PRISM and intel, so congrats on choosing an un-searchable name.
There’s also the NSA PRISM, which is older than the current LLM garbage. It’s just a basic ass name that makes for a good pun on something that changes how you see things.
Full of AI fingerprints yet no disclosure.
Yep, I’m a solo dev and I use AI assistance while building this. So, I should’ve been upfront about it. The code’s all reviewed, tested, and MIT-licensed, so it’s fully auditable. I’ll add a disclosure to the README
So you vibecoded a security product and named it after a famous government program known for spying unlawfully on American citizens
To what, capitalize on the SEO?
This is kinda gross dude, not gonna lie
