AN INTELLIGENCE SYSTEM, NOT A CHATBOT.
ORDAL investigates Web3 projects using specialized AI models and on-chain evidence, then delivers a public verdict. One project, multiple witnesses, one judge.
Resolve any input — contract, website, X account or project name — into a verified project identity.
Collect public evidence from the web, X, GitHub, docs and on-chain data. Every claim keeps its source.
Determine what kind of project this actually is, then apply category-specific scoring weights.
Grok studies social signals, Gemini studies meme and culture, OpenAI studies technology, and a program — not a model — reads the chain.
Contradictions between investigators are surfaced as conflicts, not smoothed over.
DeepSeek deliberates over the evidence store only — no new research, no invented facts — and issues the verdict.
- On-chain numbers are read by programs, never generated by language models. AI may interpret and explain that data, never invent it.
- Every investigator output is structured JSON tied to evidence IDs. The judge may only cite evidence that exists in the store.
- A failed investigator does not fail a trial — it lowers confidence and is displayed as unavailable.
- Risk signals are indicators derived from public data. They are not accusations of fraud or wrongdoing.
ORDAL analyzes publicly available information using automated systems. Results may contain errors or incomplete information. Scores and verdicts are analytical signals, not financial advice.