1. Overview
POSE measures synthetic media risk (voice, video, text) and enables parametric insurance with pre‑agreed triggers. SDKs collect signals at the edge; POSE computes a Unified Risk Score (0‑100). When a score or event threshold is exceeded, a payout is triggered. No adjusters. No disputes.
Key pillars
- Drop‑in SDKs: voice, video, text
- On‑chain Evidence NFTs (time‑stamped logs)
- Parametric payout ladder tied to risk bands
- Telemetry → dynamic premiums (like telematics)
- Treasury-backed claims with reinsurance lanes
Quick numbers (demo)
Risk score (CEO voice) | 92/100 |
---|---|
$ Prevented (YTD) | $22.8M |
Attempts blocked | 11 |
72‑hour target payout | Yes — parametric |
2. Quick start
npm i @pose/voice-sdk
import { Pose } from '@pose/voice-sdk';
const pose = new Pose({ apiKey: process.env.POSE_API_KEY });
// Browser: capture 5s audio, compute live risk
const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
const result = await pose.voice.score({ stream, domain: 'ceo-voice' });
console.log(result);
/*
{
authenticity: 0.84,
liveness: 0.90,
spoof: 0.07,
drift: 0.11,
risk: 92, // Unified Risk Score (0-100)
evidenceId: "evid_0xabc..."
}
*/
3. Claims (Parametric)
POSE uses pre‑agreed risk bands. If a logged event exceeds a threshold, a payout percentage of the limit is authorized automatically (subject to basic checks and sanctions).
Event Type | Trigger | Payout % | Evidence |
---|---|---|---|
Spoof Attempt (Low) | Risk ≥ 80 | 20% | SDK logs + Evidence NFT |
Spoof Attempt (Medium) | Risk ≥ 90 | 40% | SDK logs + Evidence NFT |
Spoof Attempt (High) | Risk ≥ 95 | 60% | SDK logs + Evidence NFT |
Confirmed Breach (Contained) | Detected & blocked | 80% | Event graph + Evidence NFT |
Confirmed Breach (Loss) | Loss incurred + POSE logs | 100% | Event graph + Evidence NFT |
How a claim flows (step‑by‑step)
- SDK posts signed telemetry → POSE computes real‑time score
- Event crosses trigger → Evidence NFT is minted (IPFS artifact hash)
- Policy engine checks ladder, sanctions, policy period
- Parametric payout authorized (bank rails or on‑chain USDC)
- Audit: everyone can verify the evidence trail
4. Tokenomics
Total supply: 1,000,000,000 Φ. The network sustains a strong claims backstop via Protocol Take and the POSE Treasury.
Allocation | Share | Vesting |
---|---|---|
Treasury (claims backstop) | 40% | Unlocked by governance |
Builders / Grants | 20% | 4y, 1y cliff |
Investors | 18% | 4y, 1y cliff |
Team | 18% | 4y, 1y cliff |
Liquidity & Market Ops | 4% | As needed |
Protocol Take: default 2.5% of premiums routed to treasury; adjustable by governance (1–5%).
Treasury model (sample)
- Annualized premium volume (APV): $120M
- Protocol Take (2.5%): $3.0M inflow
- Target backstop multiple: 3× expected annual payout
- Reinsurance lane: cedes tail beyond backstop
5. APIs
POST /v1/voice/score
POST https://api.pose.xyz/v1/voice/score
Authorization: Bearer <POSE_API_KEY>
Content-Type: application/json
{
"domain": "ceo-voice",
"audio": "<base64>"
}
200 OK
{
"authenticity": 0.84,
"liveness": 0.90,
"spoof": 0.07,
"drift": 0.11,
"risk": 92,
"evidenceId": "evid_0xabc"
}
POST /v1/claims/submit
POST https://api.pose.xyz/v1/claims/submit
Authorization: Bearer <POSE_API_KEY>
Content-Type: application/json
{
"policyId": "pol_123",
"evidenceId": "evid_0xabc",
"requested": 4700000
}
Webhook: evidence.minted
POST https://yourapp.example.com/webhooks/pose
{
"type": "evidence.minted",
"data": {
"evidenceId": "evid_0xabc",
"ipfs": "ipfs://bafy...",
"score": 92,
"domain": "ceo-voice",
"timestamp": 1732489014
}
}
6. On‑chain Evidence
Every scored incident can mint an Evidence NFT (soulbound) that anchors: model digest, device attestations, time‑series features, and time stamps. Regulators and reinsurers can verify proof‑of‑event independently.
Evidence schema (sample)
{
"schema": "pose.v1",
"digest": "sha256:9fd3...",
"features": ["mfcc","pitch","jitter","energy"],
"attestations": ["webauthn","device-id"],
"policyRef": "pol_123",
"score": 92,
"uri": "ipfs://bafy..."
}
7. Reliability & SLAs
- Evidence NFT minting: < 60s median
- Claim disposition for parametric bands: < 72 hours target
- Uptime (API): 99.95% monthly
- Data residency: US/EU selectable
8. Security
- KMS‑backed key management; HSM for production signing
- End‑to‑end TLS; token‑scoped API keys; fine‑grained roles
- PII minimization with optional on‑prem SDK scoring
- Independent audits & model cards for bias/robustness
9. FAQ
1. What is POSE?
POSE is a decentralized protocol for Proof of Self (PoS) and Proof of Voice (PoV), securing human identity in the age of AI deepfakes. It provides verified identity primitives for insurance, finance, media, and communication.
2. What is Proof of Voice (PoV)?
PoV ensures your voice is uniquely bound to you. It resists spoofing and deepfakes using advanced liveness checks, drift handling, and zero-knowledge cryptography.
3. What is Proof of Self (PoS)?
PoS extends identity verification beyond voice, combining behavior, context, and cryptographic proofs to certify “you are you” across devices and networks.
4. Why does POSE matter now?
Deepfakes and AI impersonation are exploding. Enterprises, governments, and platforms need verifiable human identity rails immediately — POSE provides them.
5. How is POSE used every day?
Users log in with their voice, sign smart contracts, verify calls, and secure online actions. Enterprises use it for fraud prevention, compliance, and claims processing.
6. How big is the market?
Identity and fraud prevention is a trillion-dollar problem. Voice biometrics alone is a $40B+ market, expanding into insurance, fintech, and social verification.
7. How many tokens exist?
Total supply is capped at 1 billion POSE tokens. Allocation covers treasury, staking rewards, ecosystem growth, and investor/partner distributions.
8. How is POSE backed?
Protocol Take (2–5% of premiums/fees) flows directly into the treasury. The treasury backs claims and provides reinsurance-like reserves, growing with usage.
9. Is POSE open source?
Yes. Core SDKs, smart contracts, and reference implementations are open source, ensuring transparency and broad developer adoption.
10. How does the network make money?
Through verification fees, protocol take from premiums, staking spreads, and marketplace activity around evidence NFTs and reinsurance markets.
11. Who are the investors?
POSE is backed by tier-1 investors including Polychain, along with strategic partners in insurance, AI, and crypto infrastructure.
12. Where is POSE based?
Founded in Palo Alto with offices in Chicago, POSE combines Silicon Valley innovation with Midwest financial/insurance expertise.
13. How is POSE going to market?
Direct integrations with insurers, banks, and enterprise apps, plus SDK/API adoption for developers building consumer and B2B apps.
14. What is the scale plan?
From pilot insurers and fintechs → to cross-industry adoption → to consumer apps and wallets embedding POSE verification globally.
15. How do investors get involved?
Investors can participate through token allocations, staking pools, and ecosystem funds aligned with network growth.
16. Will this be a top-5 token?
It can be — if POSE continues driving verifiable utility: measurable risk reduction, transparent evidence, and a treasury that genuinely backs claims.
17. How fast are payouts?
Parametric claims target instant to 72-hour settlement, depending on policy type. Treasury-backed, automated flows make it faster than traditional insurance.
18. Can I self-host the SDK?
Yes. You can run POSE SDKs locally or in your own cloud, while still connecting to the network for verification and claims.
19. Who should care about POSE?
Banks, insurers, governments, apps, developers, and consumers — anyone who needs to separate humans from bots, truth from deepfakes.
20. Does the network store voices?
No. Raw voices never leave the device. Only cryptographic fingerprints and ZK proofs are sent on-chain — protecting privacy.
21. How does POSE do zero-knowledge identity?
POSE creates zero-knowledge proofs that confirm “valid human” without exposing underlying biometric data. This protects privacy while proving authenticity.
22. Who pays for the network fees?
Typically enterprises (banks, insurers, apps) cover the fees. Consumers may pay small fees if running self-sovereign verifications.
23. How does staking work?
POSE tokens are staked by validators who run verification nodes. Stakers earn fees and rewards while securing the protocol.
24. What is the role of the treasury?
The treasury accumulates protocol take, builds reserves, backsstop claims, and funds ecosystem growth — making POSE antifragile.
25. What are Evidence NFTs?
Each claim can mint an evidence NFT containing anonymized, hashed, or redacted proof. These NFTs can be traded, staked, or audited.
26. What about regulators?
POSE is designed for compliance. Transparent claims, evidence handling, and proof standards make it regulator-friendly.
27. How does POSE prevent spoofing?
Through advanced liveness detection, anti-spoof challenges, drift monitoring, and AI-resistant voiceprint models.
28. Can POSE work offline?
Yes. POSE SDKs can run offline for local verification, syncing proofs to the blockchain once reconnected.
29. What is the adoption strategy?
Embed POSE into apps, wallets, call centers, and claims workflows. Incentivize adoption through rewards, SDK support, and ecosystem grants.
30. How does POSE integrate with Web3?
POSE plugs into wallets, DAOs, and dApps. It enables verifiable human governance, Sybil-resistant voting, and fraud-proof transactions.
31. What’s next for POSE?
Scaling identity proofs globally, building financial rails around verified humans, and moving toward becoming the universal identity layer for Web3 and beyond.
31. What’s next for POSE?
Scaling identity proofs globally, building financial rails around verified humans, and moving toward becoming the universal identity layer for Web3 and beyond.
32. Where are the voice records saved?
POSE never stores raw voice recordings on the blockchain or central servers. All audio is processed locally on-device (phone, app, or enterprise server). The output is a cryptographic fingerprint (embedding, liveness score, drift model), which is what gets verified and anchored on-chain — not the raw voice. This design ensures GDPR/HIPAA/FINRA compliance and eliminates liability.
33. How does real-time verification actually work?
When a user speaks a phrase:
- The on-device SDK generates a zero-knowledge proof in milliseconds.
- That proof is sent to POSE validators.
- Validators confirm authenticity and uniqueness.
- A verified “human attestation” is returned — typically sub-second, fast enough for live calls or logins.
34. Can POSE work during a live phone call?
Yes. On a call, POSE can verify:
- Who is speaking (identity).
- That the speech is live, not synthetic.
- What was said (if the enterprise records/transcribes — POSE handles authentication, not surveillance).
- Optional metadata like time, device, or location (if enterprise chooses).
35. Does POSE record conversations?
No. POSE never stores or records full conversations. Enterprises can record calls in their own systems if they need to, and then anchor hashes or evidence NFTs to POSE for auditability. That way, the enterprise can prove “this exact verified person said this exact thing at this exact time” without exposing raw audio.
36. Can an enterprise store everything if they want to?
Yes, but POSE doesn’t. Enterprises can store full recordings internally (standard compliance practice) and then anchor proofs to POSE. This balances:
- Privacy-first (POSE doesn’t hoard data).
- Audit-ready (evidence NFTs, hashes).
- Enterprise choice (store what’s needed, prove it via POSE).
37. How does POSE work if I’m an enterprise that records everything?
POSE fits alongside existing workflows:
- You still record everything for compliance.
- POSE SDK creates fingerprints + ZK proofs in real time.
- You hash the recording/transcript → anchor it to POSE proof.
- In a dispute, you present the raw file + POSE attestation + anchored hash = bulletproof evidence.
38. Can enterprises publish data publicly?
Yes, enterprises control the switch:
- Private-only: keep data internal, only proofs on-chain.
- Selective public anchoring: publish hashes of chosen events (e.g. paid claims).
- Evidence NFTs: tokenize redacted call logs or metadata for transparency or secondary trading.
- Full public: rare, but possible for NGOs or regulators.
39. Why would an enterprise ever publish publicly?
For trust, compliance, and capital. Public anchoring makes them audit-proof, regulator-friendly, and attractive to investors or reinsurers. It demonstrates solvency and fairness without leaking sensitive customer data.
40. Who pays if it’s public?
- Enterprises: usually pay, since transparency benefits them.
- Protocol treasury: may subsidize disclosures that strengthen the network’s trust.
- Capital markets/reinsurers: if they demand public proofs as a condition for capital, they cover the cost.
- Consumers: may pay tiny fees for self-attested public statements, but not at enterprise scale.
41. Where is the data saved?
POSE data storage follows a layered model:
- On-Device (First Layer): Raw voice/audio is processed locally on the user’s phone, browser, or enterprise server. POSE never uploads raw recordings by default. Output: cryptographic fingerprint + ZK proof.
- Blockchain (Attestations Only): The blockchain stores tiny proofs + hashes, not raw data. These attestations certify that “a human spoke this phrase at this time” — without storing the actual voice.
- Enterprise Storage (Optional): Enterprises may still record and store full calls in their own systems (often required by FINRA, HIPAA, MiFID II). They can anchor hashes of those files to POSE for tamper-proof evidence.
- IPFS / Decentralized Storage (Optional, Selective): If an enterprise or user wants public anchoring, IPFS or Arweave can be used to store redacted transcripts, evidence NFTs, or metadata. These are opt-in; by default, POSE does not push voice data to IPFS.
42. What lives on the POSE ledger? (examples)
The ledger records attestations and state changes (not raw audio): identity proofs (ZK), claims lifecycle events, treasury flows, staking/validator actions, and governance outcomes. Below are representative entries.
When | Type | Actor | Tx | Data |
---|---|---|---|---|
2025-09-25 00:01:42 | VoiceProof | 0x9aF3…c111 | 0x51de8a07… | zkProof: hash(voiceprint), liveness: 98%, drift: 0.2% |
2025-09-25 00:02:17 | SelfProof | 0x4Bb2…7ef9 | 0xa1e74ff3… | zkProof: hash(biometrics), device: iOS-Safari-16.6 |
2025-09-25 00:27:10 | ClaimFiled | 0x77Ff…a666 | 0x87697a01… | channel: Portal, insured: Acme Bank |
2025-09-25 00:27:11 | EvidenceAnchored | 0x2211…4444 | 0x6cd5bd24… | merkleRoot: 0x86fec3…, cid: ipfs://5c019a3… |
2025-09-25 00:27:11 | RiskScorePub | 0x5f6A…a555 | 0xac4d995f… | score: 92, method: Chainlink Functions |
2025-09-25 00:27:12 | AdjusterApproval | 0x77Ff…a666 | 0x5a1dbaeb… | note: Reviewed evidence and score |
2025-09-25 00:27:13 | PayoutAuthorized | 0x1b2A…b222 | 0x92a86c21… | ladder: ≥90 → 40%, cap: $1,000,000 |
2025-09-25 00:27:14 | PayoutExecuted | USDC-CCTP-Router | 0xadda227b… | amount: $640,000, route: USDC CCTP → L2 Wallet |
2025-09-25 01:02:33 | ProtocolTake | POSE-Treasury | 0x3ad6c4ff… | 2.5% fee, inflow: 25,000 USDC |
2025-09-25 01:02:40 | TreasuryBalance | POSE-Treasury | 0xf5b3a6e9… | balance: 128,450,000 POSE, 18,222,000 USDC |
2025-09-25 02:15:02 | StakeAdded | 0x33Ae…6bb7 | 0x6ab22c17… | amount: 250,000 POSE, validator: node-west-1 |
2025-09-25 02:16:11 | ValidatorReward | POSE-Rewards | 0x4d29fb88… | epoch: 2209, reward: 1,225 POSE |
2025-09-25 03:05:20 | GovProposal | 0xDao…1111 | 0x7f8bece1… | prop: Adjust protocol take from 2.5% → 3% |
2025-09-25 03:10:59 | GovVote | 0xAee1…9999 | 0x8b3e47ac… | vote: YES, weight: 120,000 POSE |
2025-09-25 03:45:04 | GovOutcome | POSE-DAO | 0xfed2913a… | result: PASSED, quorum: 72% |
2025-09-25 04:01:22 | FraudPrevented | POSE-Analytics | 0x1199fd1b… | blocked attempt: 1,232; est. loss avoided: $22.8M |
2025-09-25 04:02:12 | ProofPublished | VitalCare | 0x22feab7c… | public hash anchored: 0xa92c33…, evidence: ipfs://... |
Notes: Ledger entries are immutable attestations. Identity-related rows contain hashes/ZK proofs only—no raw audio. Claims rows capture the parametric flow end-to-end. Treasury, staking, and governance rows provide protocol transparency.
43. How does POSE compare to anything else out there?
What we’ve built with POSE is extremely strong compared to alternatives. Here’s a benchmark:
1. Transparency Layer
- Most blockchains/logging systems either stop at financial transfers (Ethereum, Solana) or audits of models/data (e.g. OpenAI eval logs).
- POSE ties multiple dimensions together in one place:
- Identity proofs (ZK voice + self)
- Claims lifecycle (parametric insurance-like)
- Treasury flows (protocol economics)
- Validator staking/rewards (consensus layer)
- Governance (DAO-style proposals/outcomes)
- ESG/fraud prevention events
→ This holistic scope is rare. Most ledgers capture only one of these domains.
2. Privacy vs Proof Balance
- Current KYC/AML systems (Jumio, Onfido, Worldcoin Orb) either:
- Store personal data → risk of breach.
- Or reduce to a single attestation with no ongoing utility.
- POSE stores hashes + zkProofs only. That’s state of the art.
- Nothing else widely deployed offers this balance: provable identity events without raw biometric leakage.
3. Composability
- Other ecosystems (e.g., Arweave + Lens, Filecoin + Ceramic) can anchor metadata.
- POSE is structured as composable event types:
- VoiceProof, SelfProof
- ClaimFiled, EvidenceAnchored, …
- GovProposal, GovVote
- etc.
- This makes it developer-friendly and auditor-friendly — much closer to Ethereum “event logs” but applied to identity and claims.
4. Enterprise-Readiness
- Insurance players like Marsh, Munich Re, or Swiss Re are experimenting with parametric claims + blockchains. But they:
- Rarely unify governance, staking, fraud prevention, and proofs.
- Lack user-side zk identity proofs.
- POSE aligns directly with enterprise compliance (FINRA, HIPAA, MiFID II anchoring), out-innovating incumbents.
5. Differentiator
Compared to:
- Worldcoin: biometric proof, but single-purpose, centralizing trust.
- Chainlink Functions / Oracles: off-chain compute proofs, but no identity ledger.
- Civic / Polygon ID: decentralized ID, but not integrated with claims, treasury, or governance.
- Insurance-on-chain (Nexus Mutual, Etherisc): parametric events, but no zk identity or fraud analytics.