Frequently asked questions

What does TRONORIGIN actually do?
It analyzes the transaction history of a TRON wallet and answers two questions. For the wallet's origin, it produces a ranked list of the addresses most likely responsible — each candidate gets a relative origin share, a confidence band (Low / Medium / High), and a plain-language explanation of why it ranked where it did. For current control, it names who can sign for the wallet now, determined from the wallet's on-chain permission evidence.
What does “origin” mean — what is the difference between activation, funding history, and current control?
Activation is who created or first funded the wallet. Funding history is who has been consistently involved over the wallet’s life — recurring transfers (including the small TRX top-ups that keep a wallet fueled) and resource delegation all count. Current control is who can sign for the wallet right now. Activation and funding history together answer the origin question and are scored into relative origin shares. Current control is answered separately — rule-first from the wallet's on-chain permission keys and permission-change history, not from a score — because it is a different question, and the two answers can legitimately disagree.
How does the confidence score work?
The origin score weighs activation and funding-history signals (transaction patterns, resource delegation, return flows, sustained USD-valued funding, and more), applies bonuses and penalties, and produces normalized origin shares. The origin confidence band is then set by a calibrated probability that the top candidate is the wallet’s true activator: higher calibrated probability maps to Medium and then High. That calibration still rests on the underlying evidence — the strength of the top candidate’s signals, how dominant its origin share is, and whether the activation-window and longer-term funding evidence corroborate each other — while the gap to the next-closest candidate drives the separate stability indicator. The controller headline carries its own confidence, based on how direct the permission evidence is.
What is the calibrated activation probability?
Alongside the Low / Medium / High band, an analysis can report a calibrated probability that the top candidate is the wallet’s activator — a point estimate with a range around it. It is an input to the confidence band, not the same thing: the band is a coarse label, while this figure expresses the model’s calibrated certainty more precisely. It appears when probability calibration is enabled for the analysis.
Where does TRONORIGIN get its data?
Analysis draws on public TRON blockchain data via TRONscan and TronGrid APIs, together with Bitquery, a full-history archive source, so the complete on-chain record is available. Token price lookups use CoinGecko. Security flags (blacklist status, fraud history) come from TRONscan’s security APIs. All data sources are public; TRONORIGIN does not have access to any private or off-chain information.
What is takeover detection?
Takeover detection fires when the address that dominated the wallet’s early history is different from the address that dominates its recent activity, and the divergence is significant enough to be unlikely by chance. When this happens, TRONORIGIN flags the result and classifies the likely type of transition — for example, a sale, a wallet compromise, or a secondary wallet coming into regular use. It does not mean something malicious happened; it means the data suggests control changed hands.
Can it identify a specific person?
No. TRONORIGIN attributes addresses to known clusters, entities, or labeled addresses (exchanges, contracts, faucets, and so on) where the on-chain evidence supports that attribution. It does not attempt to link blockchain addresses to real-world identities. If a candidate address is already labeled in the registry — for example, as a Binance hot wallet — that label will appear; otherwise, you see a raw address.
How accurate is it?
Accuracy depends on the wallet’s history. Wallets with rich, consistent activity and clear delegation relationships tend to produce reliable results. Wallets with very few transactions, heavy mixer involvement, or deliberately obfuscated histories will show lower confidence and may not produce actionable attribution. The confidence band and data quality indicator (Insufficient / Limited / Adequate / Rich) are there to tell you how much weight to put on any given result. Because the analysis reads a wallet’s complete on-chain history, thin results reflect a genuinely sparse record rather than a truncated sample.
Is it free?
Yes, the tool is free to use. Running a standard or deep-scan analysis costs nothing. If usage patterns change in the future, pricing information will be posted here.
Is anything sent to a server when I analyze an address?
When you submit an address, the request goes to a stateless serverless analysis function, which fans out to public TRON data sources (TRONscan, TronGrid, Bitquery, CoinGecko) to fetch transaction history and account data. The AI summary feature sends the structured analysis result to the Anthropic Claude API to generate the natural-language explanation. No query history is stored on TRONORIGIN’s infrastructure; each analysis is stateless. Be aware that requests to third-party APIs (TRONscan, TronGrid, Bitquery, CoinGecko, Anthropic) are subject to those providers’ own data handling practices.
Why does the AI summary sometimes hedge?
The AI summary is generated from the structured analysis result — it reflects whatever confidence level the heuristic produced. When the underlying data is ambiguous, limited, or contradicted by competing signals, the summary will say so rather than projecting false certainty. That hedging is intentional: the goal is to help investigators make informed decisions, not to give them a confident-sounding answer that the data doesn’t support.
I found a wrong attribution — what should I do?
Use the feedback button in the analysis view to report it. Include the address you analyzed and a brief description of what you believe the correct attribution is. Attribution errors are the most valuable feedback the project receives and directly inform scoring improvements.