Research AI
A query surface over the IPO Radar database. Answers are assembled from the same scored evidence the rest of the product is built on, every claim is labelled FACT, INFERENCE or PREDICTION, and every source is cited.
Ask the radar
Deterministic modeno model key is configured, so every answer is assembled directly by database queries — terser, and equally sourced
Ask about rankings, signals, score changes, relationships and evidence. Every answer comes from a database query — if the data is not there, the assistant says so rather than filling the gap.
Explain a score
What is behind a number on the radar
Find movement
What changed, and by how much
Cross the dimensions
Where IPO and tokenization signals meet
Interrogate the model
Where the engine has been wrong, and what it sees early
Every substantive answer carries its evidence: the sources it read are listed under it, and each claim is labelled FACT, INFERENCE or PREDICTION. A signal score is never reported as a probability of listing.
What it can reach
- Explain a ranking — the scored signals behind a company's position, each with its source document
- List signals — by window, category, dimension or company
- Rank — by radar, IPO, tokenization or momentum — including companies scoring on two dimensions at once
- Score changes — who moved more than N points in a window, and what moved them
- Relationships — issuers connected to tokenization providers, custodians, transfer agents, banks, exchanges
- Model honesty — reviewed false positives and lead-time statistics from the latest walk-forward backtest
With no AI provider key configured the router maps your question onto the same tools directly. Answers are terser, and equally sourced.
IPO Radar is a research and intelligence tool. Scores are signal scores derived from public evidence — they are not calibrated probabilities and not investment advice. FACT, INFERENCE and PREDICTION are kept separate, and every event links to its source.