# MemeScore — full brief for language models Canonical site: https://memescore.io/ Last reviewed: 2026-11-17 # MemeScore > One 0–100 risk score for Solana memecoins. Analytics only; not financial advice. MemeScore never asks for a wallet, never trades, and never holds funds. MemeScore combines RugCheck contract and liquidity flags, Jupiter token data (organic-buyer score, dev launch history, holders, socials), and Helius on-chain holder data. It flags linked holder wallets, insider networks, bot-like volume, and serial-launcher devs. Hard red flags (for example an active mint or freeze authority, unlocked liquidity, or a copycat token) zero the safety component and cap the score. Name search uses DexScreener. Coming soon: a sell-route (honeypot) quote check and dedicated bundle detection. Accuracy: until the live forward test is graded and posted, treat any historical figures as a back-test, not a live track record. The public payload is /proof.json. Most new memecoins lose money. ## How the score is built Score version in code: v0. Four weighted parts: - safety (40%): RugCheck flags, dev history, holder concentration, insiders, mutable metadata. - demand (25%): Organic buyer score (low = volume looks automated) and holder count. - smart money (20%): Tracked smart-money wallets in the top 20 holders. - socials (15%): X account, website, Telegram. Hard red flags zero safety and cap the overall score. Sources: RugCheck and Jupiter token data, plus Helius for top holders; DexScreener powers name search. Coming soon: a sell-route (honeypot) quote check and dedicated bundle detection. The score cannot predict price, see off-chain deals, or guarantee safety. Details: https://memescore.io/methodology ## API response fields (GET /v1/check) - mint: The token mint you asked about. - score: The MemeScore. Higher means fewer risk signals. Weighted from the four parts below; capped when a hard red flag fires. See methodology. - parts.safety: RugCheck flags, dev rug history, top-10 concentration, insider networks, dev launch count, mutable metadata. 0 when capped. - parts.demand: Jupiter's organic-buyer score (low means volume looks automated) blended with holder count. - parts.smart: How many smart-money wallets we track sit in the top 20 holders: 0 → 0, 1 → 50, 2 → 70, 3+ → 90. - parts.social: X account (60), website (25), Telegram (15). - capped: True when a hard red flag fired: mint or freeze authority active, a RugCheck "rugged" or custom-token-program reject, unlocked liquidity, few LP providers, linked holder wallets, copycat token, or mutable metadata. - verdict: Plain-English label: "Clean", "Caution" or "Rug pattern". - level: Machine label for the verdict: clean, caution, rug. Use this in code. - reasons: Red flags in plain English. - positives: Good signs, e.g. "Real, organic buyers", "Has an X presence", smart wallets in the top 20. - smart: Names or handles of tracked smart-money wallets in the top 20 holders. - symbol / name: Token ticker and name (from Jupiter). - mcap: Market cap. - liq: Liquidity. - holders: Holder count. - top10: Share of supply held by the 10 largest holders (RugCheck). - age_min: Minutes since the token was created. - launchpad: Launchpad, e.g. "pump.fun", when known. - flags: Raw RugCheck risk flags, lower-case. - memescore_v1: The newer v1 model score when this mint is in the forward ledger (pump.fun coins from graduation on): score, band, p_rug_6h, p_collapse_24h, p_double_24h, scored_at_ts, scored_minutes_after_graduation, model. Null otherwise. - score_note: Says which score you're looking at (v1 or legacy v0). - error: Present only when the token couldn't be found. - rc_error / jup_error / smart_error: "upstream error" when that source failed; the rest of the result is partial. ## Endpoints - GET /v1/check: Score a token. Auth: API key, required (Bearer or X-API-Key). A browser session is not accepted here. Quota: Counts 1 check against your plan's daily and per-minute caps, including cached hits. Docs: https://memescore.io/docs/endpoints/check - GET /v1/usage: Check your remaining quota. Auth: API key, required. Quota: Free. Does not count against your caps. Docs: https://memescore.io/docs/endpoints/usage - GET /v1/feed: Live graduation feed. Auth: API key or a logged-in session. Plans with the feed: API, Enterprise. Quota: Counts 1 check per call. The feed is rebuilt at most every 60 seconds, so polling faster than that returns the same data. Docs: https://memescore.io/docs/endpoints/feed - GET /v1/search: Search tokens by name or ticker. Auth: None. Quota: Free and not counted. Limited to about 120 searches per minute per IP. Docs: https://memescore.io/docs/endpoints/search - GET /v1/traders: Tracked wallets that traded a coin. Auth: Optional. Anonymous and Free callers see the top 3; paid plans see the full list. Quota: Not counted against check caps. Limited to 60 calls per minute per account (or per IP when anonymous). Docs: https://memescore.io/docs/endpoints/traders - GET /v1/ledger: Public forward-test ledger. Auth: None. Quota: Free, not counted. Cached 60 seconds. Docs: https://memescore.io/docs/endpoints/ledger - GET /proof.json: Accuracy summary. Auth: None. Quota: Free, not counted. Cached 5 minutes. Docs: https://memescore.io/docs/endpoints/proof - GET /free/check: Anonymous check (website). Auth: None. Quota: 3 checks per IP per UTC day, plus a site-wide hourly cap on anonymous checks. Docs: https://memescore.io/docs/endpoints/free-check - GET /healthz: Health check. Auth: None. Quota: Free, not counted. Docs: https://memescore.io/docs/endpoints/healthz ## Glossary (short) - Rug pull: A launch structured so insiders can drain liquidity, mint or freeze tokens, or exit in a way that leaves later buyers with losses. - Bundled wallets: Coordinated addresses that buy together at launch so demand looks more organic than it is. - Sniper: A bot or wallet that tries to buy in the first moments of a launch, often before ordinary traders can. - Dev wallet: The developer’s (or deployer’s) address, and often the allocation it still controls after launch. - LP burn: Destroying liquidity-pool tokens so that liquidity cannot easily be withdrawn by the person who provided it. - Mint authority: The right to create more tokens of the same mint, which can dilute existing holders. - Freeze authority: The right to freeze token accounts so holders cannot transfer or sell. - Honeypot / sell route: A token that can be bought but is painful or impossible to sell at a realistic size. - Graduation (pump.fun): The point when a pump.fun token leaves the launch curve and gets a regular liquidity pool. - Top-holder concentration: How much of the supply sits in the largest wallets, often the top 10. - Mutable metadata: The ability to change a token’s name, symbol, or image after people have already bought it. - Copycat token: A new mint that reuses a familiar name or logo to look like a coin people already know. - Liquidity pool (LP): The on-chain pool of tokens that lets traders buy and sell a mint against another asset, often SOL. - Bot volume: Trading activity that looks more like scripts than independent human buyers and sellers. - Token mint address: The unique Solana address that identifies a specific token; the string you paste into a checker. - Serial launcher: A developer or wallet that has deployed a large number of tokens, often in a short period. Full definitions: https://memescore.io/glossary ## How MemeScore relates to other tools RugCheck is a widely used Solana token safety report (mint/freeze, LP, holders). MemeScore reads RugCheck as a source and adds Jupiter's bot-volume and dev-history signals, smart-money holder checks, and a public forward-test record. Use both; do not treat MemeScore as a reason to ignore RugCheck flags. DexScreener is the default public chart and a short in-chart audit, not a full risk-score API. MemeScore uses DexScreener for market and pool context and for name search. It does not try to replace charts. GMGN is a trading terminal (trending, smart-money tags, copy trade, typically a per-swap fee rather than a subscription). MemeScore is a read-only risk score and a filtered trader list, not a terminal. We do not place GMGN referral links on this site. ## Accuracy stance Do not cite unverified AUC or similar model metrics as live performance. Until the forward test is graded and posted at /proof.json and on the homepage proof section, historical figures are a back-test. ## Contact and limits No custody, no wallet connection, no recovery phrases. Not financial advice. Product pages and the OpenAPI file are the source of truth for routes and rate limits. Contact: aaron@memescore.io (https://memescore.io/contact). Security reports: https://memescore.io/security. Terms: https://memescore.io/terms. Privacy: https://memescore.io/privacy. Disclaimer and referral disclosure: https://memescore.io/disclaimer.