Foundry in 60 seconds —no wallet required.
You can verify every claim Foundry makes from this page. The inference widget calls a live Ingot on 0G Aristotle mainnet. The deployment table links to aristotle. The dashboard counters are rendered from on-chain events. Nothing here is staged.
Call a co-owned model.Right now.
The request below routes through the OpenAI-compatible Foundry proxy to a live Ingot on 0G Aristotle mainnet. The inference fee is reserved on-chain via the 0G Compute serving broker, and a portion deposits into the Ingot's RevenueSplitter — claimable by its co-owners.
- Ingot:
0x8e2af4a0…0001 - Adapter: OpenAI-compat · also via Vercel AI SDK or LangChain
- Revenue routed: pro-rata to share-ledger holders, pull-claim on chain
POST /api/v1/chat/completionsReproduce with curl
curl https://foundryprotocol.xyz/api/v1/chat/completions \
-H "content-type: application/json" \
-H "x-foundry-ingot-id: 0x8e2af4a000000000000000000000000000000001" \
-d '{"messages":[{"role":"user","content":"Tum kasso assa?"}]}'The five stops
Each stop verifies a different claim. Total time: about ten minutes end-to-end. Two minutes if you only want stop 1.
- 01
Run inference against a live Ingot
Use the widget below. Request routes through the OpenAI-compatible Foundry proxy → 0G Compute via the serving broker. The response carries a mode flag (live | tee | stub) and on-chain tx hashes when fees are reserved.
Jump to widget → - 02
Inspect a Forge end-to-end
The Forge explorer shows every contribution, the eval state, the TEE attestation, and the resulting Ingot's cap table. State transitions are emitted from on-chain events — nothing is faked.
Open Forge explorer → - 03
Trace revenue routing on-chain
Every inference call deposits into RevenueSplitter.receivePayment(tokenId). Co-owners pull-claim. The Ingot page shows claimable balance per address; the dashboard shows protocol-wide totals.
Open dashboard → - 04
Read the SDK in 3 lines
The npm package wraps everything — three lines of TypeScript get you a working inference call with revenue routed automatically. Vercel AI / LangChain / OpenAI-compat adapters slot Ingots into existing agent code.
See the SDK → - 05
Walk the lineage graph
Every Ingot has a parent Forge; reforged Ingots have parent Ingots. The graph is the public record of how every model on Foundry was made.
Open lineage graph →
Verifiable deployment
All six contracts live on 0G Aristotle (chain id 16661). Source is verified on the 0G explorer. Click any address to inspect.
| Contract | Address | Role |
|---|---|---|
| FORGEToken | 0xE716B0260f462b2A1789cB6cfCBd825736b920Ca | |
| ForgeFactory | 0x636109264EBF6cFD18CC38bD43eDf9cCad7ae23D | |
| ContributionRegistry | 0x05235Ba0F2a77bcaB87371E4d797D6830ddC2d86 | |
| Ingot | 0x39B736f424754d05a0da186d89015b74d1DDe1d3 | |
| RevenueSplitter | 0xC58E0F32BD43e43153D3CA8ee8F25C8198789289 | |
| IngotRegistry | 0xF8f3fAE648A8d7ee4Df0A7b10a0F759938aab7e1 |
What we want you to grade us on
No other project on 0G — or in the broader AI×crypto landscape — ships a primitive that turns “contributing to a model’s training” into “owning a share of its revenue forever.” See protocol overview.
Six contracts, ~100% line coverage, TEE attestation parsed on-chain, npm SDK with three adapters, Next.js 16 web surface, indexer, eval coordinator. None of it is mocked. See real vs roadmap.
Every 0G surface is load-bearing. Foundry generates chain txs, storage uploads, compute hours, TEE jobs, and Agent IDs as the direct result of its core loop. Every Forge created grows 0G. See integration matrix.
The hackathon ends; Foundry doesn’t. Audit, Shapley v2, secondary market for shares, and 50 Forges live by Month 6 are all funded by current grant runway. See the vision doc.
Live (from on-chain events): 12 Forges · 10 Ingots · 32 contributions · last block 37517010