Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
Airdrop Checker
Live surface: three.ws/airdrops
Paste any Solana or Ethereum wallet and see which tracked airdrop programs it qualifies for. The score comes from a real scan of the wallet's on-chain history, the checklist says exactly what is missing and what to do next, and every report is deep-linkable. Free, keyless for Solana, no account.
How it works
One endpoint does everything:
GET /api/crypto/airdrops # the tracked-program directory
GET /api/crypto/airdrops?address=<wallet> # scan + score that wallet
Three layers behind it:
- Activity scan:
api/_lib/wallet-activity.js. Solana wallets are scanned keylessly: transaction signatures over the platform's rotating RPC chain (count, distinct active days, wallet age, recency) plus token diversity from the shared balance layer. EVM wallets are scanned across Ethereum, Optimism, Base and Arbitrum through Etherscan's V2 unified API, where oneETHERSCAN_API_KEYcovers every chain; native volume is converted to USD before any USD threshold sees it. Scans cache for an hour per wallet. - Scoring:
api/_lib/airdrop-eligibility.js, a pure evaluator. Each registry criterion with acheckstring ("tx_count >= 50") is compared against the measured activity; the score is the share of measurable criteria met. At or above 80 the wallet is qualified, at or above 30 in progress, below that not eligible. - Registry:
data/airdrops.json, the tracked programs with their criteria, status (confirmed / upcoming / speculation), estimated ranges, and source links. It is a plain JSON file, dated, and expected to be edited as programs confirm or close.
The honesty contract
- Fields a scan cannot measure are null, never zero. A criterion over an unmeasured field is reported as unknown and counts as unmet; the wallet is told what could not be seen, not given credit for it.
- Manual steps are not scored. Protocol-specific actions (staking in a
program, providing LP) cannot be seen by a generic wallet scan, so registry
criteria without a
checkare surfaced as "do this yourself" to-dos and excluded from the score. The score only ever claims what was measured. - Capped scans say so. Signature and transaction pagination is bounded; a wallet past the cap gets flagged totals that read as honest minimums.
- Estimated ranges are labeled speculation. The hero sum only adds ranges from programs the wallet actually qualifies for, and the page repeats that these are public speculation, not promises. Nothing on the surface is an endorsement of any program.
Response shape (scan)
{
"address": "…",
"family": "solana",
"ts": "2026-08-06T20:00:00.000Z",
"activity": {
"family": "solana",
"tx_count": 342, "days_active": 61, "account_age_days": 410,
"last_active_days": 1, "unique_tokens": 12, "chains_active": 1,
"contract_interactions": null, "volume_usd": null,
"capped": false, "chains": ["solana"]
},
"opportunities": [
{
"id": "…", "name": "…", "chain": "solana", "status": "confirmed",
"score": 75, "eligibility": "in_progress",
"met": [{ "description": "50+ transactions on Solana" }],
"missing": [{ "description": "…", "recommendation": "…" }],
"manual": [{ "description": "…" }],
"estimatedValue": "$100 - $2,000", "source": "https://…"
}
],
"otherFamily": [],
"summary": { "tracked": 6, "qualified": 2, "in_progress": 3, "not_eligible": 1, "estimatedValue": { "lo": 150, "hi": 3500, "entries": 2 } },
"thresholds": { "qualified": 80, "inProgress": 30 },
"registryUpdated": "2026-08-06"
}
Registry entries for the other chain family come back in otherFamily,
unevaluated, so the page can say "check these with your other wallet" instead
of rendering a fake zero.
Related surfaces
/portfolio: the live wallet portfolio; the two pages cross-link and share the recent-wallet list./api/crypto/wallet: the raw balance endpoint the token diversity measurement rides on.