About

Foundry is a professional space for ideas in the making.

It's where research groups, study cohorts, and small teams gather around the topics that matter to them — aerospace, biotech, computer science, and the work that doesn't fit a category yet.

Why we built it

Anonymous communities are great for venting and gossip. They're a worse place for the kind of work that benefits from continuity — knowing who you're collaborating with, what they've contributed before, and whether their answer is worth trusting.

Foundry is that other kind of community. It uses real names, real headshots, and real affiliations — pulled straight from your Google account at sign-in. The bar to enter is low; the bar to misbehave is high.

Real identities, the Facebook way

Signup is one click. Your Google name becomes your Foundry name — locked, no anonymous handles. From then on, the system quietly watches for fake-identity signals: heuristics on names, mass reports, suspicious activity. When something trips, the account is flagged and asked to verify with any plausible document — a utility bill, university card, magazine subscription, ID — that confirms the name on file. Vertex AI reads it; an admin reviews edge cases.

Honest accounts never notice the system exists. Bad actors hit a wall quickly.

Topics → Communities → Teams → Threads

  • Topics are tags that anchor the directory. Seeded defaults exist, and anyone can create a new topic when starting a community or team.
  • Communities are open spaces under a topic. Anyone with a verified Foundry account can join.
  • Teams are private working groups under the same topics. Invite-only, threads stay with members.
  • Threads are long-form discussions. When the conversation branches, carve out a named subthread to keep it organised.

Community AI sharing — the noble act

Tag a thread with @communityand it lands in the Foundry query pool. Plus & Pro members can then lend their AI agent to answer it — billed at one-third the usual tokens. Metantel covers the rest.

The discount is the point. Sharing what you have with the community is a genuinely noble act, and we want to make it the most affordable way to use your AI. Answers are grounded in the community's own past threads via the shared RAG service, so they build on what the group already knows.

For students specifically, mark a thread as a tutorial sheet. AI replies on tutorial sheets get a validation badge, and peer votes become the signal for which AI answers actually helped.

What's next

Foundry is one half of a pair. When a thread is ready to become a real project — with docs, a timeline, tasks, a whiteboard, and coding-agent context — it spins into Artel, the AI-native collaborative workspace that ships next.