Token Investigation Tools

Track a CA (combined)

One click, does everything currently automatable for a single token: (1) a targeted PumpFun full-history scan for THIS token specifically, and (2) one general relayer sync tick. Important asymmetry: PumpFun can genuinely reach "complete" for this token by clicking repeatedly. FOMO can't be targeted the same way — it only reports whatever the global, token-agnostic relayer scanner has found for this mint so far, since FOMO activity is only discoverable by walking the relayer's own timeline, not by token.

If pumpfun.fullyCaughtUp is false, click again — it resumes exactly where it left off for this token. There's no equivalent "done" state for the fomo section by design (see note above).

Background relayer sync

Runs automatically once a day (the most Vercel's free plan allows). Click below to run it immediately instead of waiting — this walks the relayer's own transaction history and records every trade it's found, for every token, into persistent storage. Requires Redis (Upstash) to be connected to this project — see the README for setup, since Vercel KV was discontinued.

Full-history token scan (PumpFun)

Unlike the main dashboard's 24H panel (a live, time-boxed scan recomputed on every visit), this permanently records a SPECIFIC token's entire PumpFun history — appropriate for a low-market-cap token with a small enough total history to fully scan in one or a few clicks. Click repeatedly (or wire this into a scheduler like the sync above) until the result says fullyCaughtUp: true. FOMO-side history doesn't need this — the relayer sync above already covers every token automatically once it's caught up.

Discover additional relayers

Checks whether FOMO uses more than one gas-sponsor address. Samples up to 15 of the most recently active wallets from the permanent known-traders registry (now 90+ real confirmed FOMO wallets, no manual handle-typing needed), and looks for any OTHER address — besides the one already configured — that pays fees for multiple of them. Same underlying signal as the original manual investigation, now automated against a much bigger, real sample.

Fix FOMO sells in already-stored data

Before the sell fix, every FOMO sell was saved as a "buy" by the token's pool / bonding curve. This re-checks a token's saved trades against the real transactions, and removes pool accounts from the saved traders and buyers lists (only addresses the chain confirms are pools). Leave the box empty to run just the traders-list cleanup. If it says "Not finished yet", click again.

Re-value already-stored FOMO transactions

The volume-valuation fix only applies to transactions recorded from now on — anything already stored (like the $0.24-of-32 case) has its old, mostly-null values baked in. This re-fetches each already-stored unvalued transaction for a specific token and re-runs the corrected logic against it. Processes up to 100 at a time — click again if it reports more remaining.

Debug: raw Helius asset data

Dumps the COMPLETE, untyped Helius getAsset response for a mint — no narrowing, no assumptions about field names. Used to find real bugs (like content.files[] existing but never being read) instead of guessing at the next field name.

Debug: raw Helius transaction data

Dumps the COMPLETE, untyped Helius parsed-transaction response for one signature — including events.swap, which our own classification code has never read. Get a real signature from fomo.family's swap feed or a block explorer link, paste it here.

Debug: exact stored tracked-token record

Dumps the EXACT record currently stored for a mint in the tracked-tokens list — ground truth on what's actually saved, not a guess based on reading the code.

Debug: raw FOMO per-transaction classification

Sells showing 0 across multiple different tokens is suspicious enough to check for real — this dumps every stored classified record for a mint (direction, trader, signature), so we can see whether "sell" is ever actually produced, instead of guessing at the aggregation code again.

Debug: live-path FOMO classification

The store-based debug tool above only checks persisted data — for a token the background scanner hasn't reached yet, the live numbers on the dashboard come from a DIFFERENT code path. This calls that exact live path directly and returns every classified record, confirmed against real sell activity FOMO's own site shows for this token.