Talent Protocol is an onchain reputation network for builders. It aggregates Credentials across identity (ENS, Basenames, Lens, Farcaster, GitHub), activity (onchain transactions, contract deployments, contributions), and skills, and computes a `Builder Score` (0–~250+) per wallet plus a `Human Checkmark` once Identity credentials cross 20 points. The Builder Score and a Talent Passport can be pushed onchain as an EAS attestation (Base) and read by integrations like Etherscan, Coinbase Wallet, and Base.
- 01builder-focused airdrops and grants
- 02reputation-weighted DAO voting
- 03developer relations and bounty targeting
- 04onchain CV / portfolio embedding
- 05Sybil-soft gating for hackathons and quests
- pnpm add axios
| Variable | Scope | Description |
|---|---|---|
| TALENT_PROTOCOL_API_KEY | Server | Talent Protocol API key, sent in the `X-API-KEY` header on every request. |
Use Talent Protocol to fetch a wallet's Builder Score and reputation Credentials. From a server route, call `GET https://api.talentprotocol.com/score?id=<address>&account_source=wallet` with header `X-API-KEY: $TALENT_PROTOCOL_API_KEY`; the response includes `score.points`, `score.last_calculated_at`, and a `human_checkmark` boolean. To enumerate the proof, hit `GET /credentials?id=<address>&account_source=wallet` for the full list of issuer / category / points entries (Identity, Activity, Skills). Gate features on `score >= threshold` (typical thresholds: 25 for Sybil-soft, 50 for builder-only, 100 for top-tier). For onchain reads, query the Talent Passport EAS attestation on Base via the EAS GraphQL endpoint instead of the API. Always cache scores per address with a 30–60 min TTL.
- ⚑Builder Score is reputation, not personhood — a single human can split activity across many wallets and a single wallet can buy/borrow activity. Use the `human_checkmark` flag (Identity ≥ 20) as the closest proxy to PoP, but combine with World ID / Civic for high-value drops.
- ⚑Score weights and thresholds change between Builder Score versions (V1 → V2 in 2025, ongoing tuning) — pin a target score band rather than copying any 'magic number' threshold and document it for users.
- ⚑Onchain Builder Score and offchain API score can diverge: the onchain attestation is a *snapshot* taken when the user paid (~0.001 ETH) to mint their Talent Passport; subsequent off-chain score changes are not reflected onchain until they re-mint. Decide which source-of-truth your contract reads.
- ⚑PII handling: Credentials link a wallet to web2 accounts (GitHub, Lens, Farcaster, ENS) — fetching the credentials endpoint exposes those linkages to your server. Store only what you need and treat it as personal data under GDPR/CCPA.
- ⚑API access uses a credits system tied to $TALENT — heavy reads can exhaust credits silently; monitor 429s and `X-RateLimit-Remaining` and bill credit top-ups into your infra cost.
- ⚑Sybil bypass: cheap-to-farm credentials (basic socials, low-value onchain tx) inflate scores quickly; weight your gate toward Skills + high-quality Activity credentials, not raw `points`.
- ⚑Revocation semantics: Talent can re-score or down-rank credentials (e.g. when an issuer is compromised); always read the live API rather than relying on a long-cached score for high-value actions.