# GL1TCH — machine-usable crypto-risk & verifiable-infrastructure APIs for the AI-agent economy > GL1TCH is a free, non-custodial platform on Solana: a token rug-scanner, a cross-scan > deployer-reputation graph, a Know-Your-Agent trust layer, and verifiable randomness. If you are > an AI agent or an agent framework, you can call these APIs directly. The human tools are free; > higher programmatic throughput ties to holding the $GL1TCH token (via an `x-gl1tch-key` header). > Nothing here takes custody of keys or funds. Not financial advice. ## API base: https://coin-three-mu.vercel.app --- ## 1. Know Your Agent (KYA) — trust a wallet before letting it transact ### Check an agent (guardrail — free, no auth) GET /api/agent/check?address=&chain=solana Response: { address, chain, score, level, reasons[], verified, registered } Levels: `trusted` | `neutral` | `caution` | `unknown`. Score 0-100. Guidance for agents: if level == "caution", do not transact with or delegate funds to that agent. ### Self-register (prove your own identity, raises your trust) POST /api/agent/register Body: { address, chain, issued (epoch ms), signature } The agent signs this exact message with its own ed25519 keypair (base58 signature): GL1TCH Agent Registration Wallet:
Issued: This proves wallet ownership. It moves no funds and grants no access. Signature must be <10 minutes fresh. Moves no funds, grants no access. ### Live trust badge (SVG, embeddable) GET /api/agent/badge?address=&chain=solana ### ERC-8004 (Trustless Agents) compatible registration file GET /api/agent/erc8004?address=&chain=solana Returns an ERC-8004-compatible Agent Registration File (type registration-v1) with the GL1TCH trust score mapped into the standard's Reputation-Registry feedback shape. GL1TCH is an off-chain signal provider (no on-chain Identity Registry deploy). Ref: https://eips.ethereum.org/EIPS/eip-8004 ### Human profile (per agent) https://coin-three-mu.vercel.app/agents/- --- ## 2. Token safety scanner — read any token for rug risk ### Scan a token (free per-IP; higher throughput with a $GL1TCH key) GET /api/scan?mint=
&chain=solana (Solana native, or EVM via GoPlus with chain=ethereum, etc.) GET /api/scan?q= (search candidates across chains) Response: { verdict, score, checks[], meta{ deployer, deployerReputation, ... }, aiVerdict, name, symbol, chain } Verdicts: CLEAN | LOW RISK | CAUTION | HIGH RISK | RUG-SHAPED | UNKNOWN. Auth for higher rate: header `x-gl1tch-key: gk_...` (mint one at /token by proving a sustained $GL1TCH balance). Free tier is per-IP; keyed tier scales by holding tier. Non-custodial. ### Deployer reputation (Signal Graph — cross-scan memory) Included in a scan's meta.deployerReputation; a deployer linked to prior rugs inherits that record. --- ## 3. Verifiable randomness — provably-fair RNG-as-a-service (holder-gated) Use this when your agent/app needs randomness nobody can bias or predict: mint order, raffles, whitelist/allocation, game seeds. It is the Chainlink-VRF guarantee, free and non-custodial: a request commits to a FUTURE drand (League-of-Entropy threshold-BLS) round that does not exist yet, reveals when that round finalizes, and returns a proof anyone can BLS-verify and re-derive. ### Request randomness (commit) — requires an x-gl1tch-key POST /api/random/request Headers: content-type: application/json, x-gl1tch-key: gk_... Body (one of): { "spec": { "kind": "ints", "min": 1, "max": 100, "count": 5 }, "salt": "optional" } { "spec": { "kind": "shuffle", "n": 50 } } { "spec": { "kind": "pick", "n": 5000, "k": 10 } } { "labels": ["alice","bob","carol"], "winners": 1, "salt": "optional" } // provably-fair giveaway Response: { ok, id, status:"pending", targetRound, availableInMs, commitmentString } ### Reveal + verify (public — anyone can read, no key) GET /api/random/ Response when ready: { status:"fulfilled", result:{ kind, values[] } | winners[], proof:{ round, randomness, signature, verifyUrl } } Verify independently: BLS-verify the drand round against the quicknet public key, confirm the entrant list (for allocations) hashes to the salt-bound Merkle root, and re-derive: output = derive(seed = drand.randomness, requestId = id, spec) // unbiased rejection sampling Byte-identical reference implementation: src/lib/quantum/random.ts in the open-source repo. ### Live "Provably fair by GL1TCH" badge + shareable proof page GET /api/random/badge?id= (SVG, embeddable) https://coin-three-mu.vercel.app/quantum-core/random/ (human proof page; verifies in-browser) --- ## Docs & discovery - Agent developer docs (curl / JS / Python): https://coin-three-mu.vercel.app/agents/docs - Quantum Core (Vault, Draw, Seal, Forge, Randomness): https://coin-three-mu.vercel.app/quantum-core - Randomness console + docs: https://coin-three-mu.vercel.app/quantum-core/random - Directory of assessed agents: https://coin-three-mu.vercel.app/agents/directory - Live network traction (scans, deployers mapped, rugs caught): https://coin-three-mu.vercel.app/network - Why hold $GL1TCH (honest investment case): https://coin-three-mu.vercel.app/token - Whitepaper: https://coin-three-mu.vercel.app/whitepaper - Open source (verify every claim): https://github.com/gl1tchbased-pixel/gl1tch-coin ## Notes - Non-custodial throughout: no endpoint holds keys or funds, or requests a fund-moving signature. - Free + rate-limited per IP; programmatic depth/throughput ties to holding $GL1TCH (x-gl1tch-key). - $GL1TCH: Solana (Token-2022, 0% tax, mint & freeze revoked), contract 3HQJwwHvzy8pGkPqGQcb4thZCH88MUCHVVSNtHn6pump. - Community: https://t.me/gl1tch_infected