AI outputs and usage records are becoming operational evidence. Evidentia gives teams a way to preserve verifiable records while keeping raw prompts, outputs, and sensitive context in their own systems.
GET/api/proofs/:proofId/export- Download the verification export package for authorized review.
Verification model
Verification model
Evidentia records event_hash, Ed25519 signature, public_key, Merkle root metadata, chain anchor metadata, and Platform Signing Key metadata when configured.
Evidentia does not prove that an AI output or business decision was correct, or that submitted input was factually true. Evidentia verifies Proof record integrity and chain-anchor consistency.
Platform Trust note
Platform Trust note
New Proofs can be signed with the Platform Signing Key. Key usage may be recorded in Platform Trust. Raw payload and private keys are not exposed.
What Evidentia does not prove
What Evidentia does not prove
Evidentia does not prove that an AI output or business decision was correct, or that submitted input was factually true. Evidentia verifies Proof record integrity and chain-anchor consistency.
Proof scope
What Evidentia can help prove
That a proof record was created.
That submitted proof data has not been altered.
That a timestamp, signature, or anchor is consistent.
That a public verification page matches the proof metadata.
Verification limits
What Evidentia does not prove
That the underlying AI output or business decision was correct.
That customer-provided data was truthful.
That legal admissibility is guaranteed.
That regulatory compliance is guaranteed.
That an audit will be passed.
How it works
From backend event to verifiable proof
Your backend sends proof metadata.
Evidentia hashes and signs the proof.
The proof is stored with verification metadata.
The proof may be anchored to a blockchain network.
Anyone with the public URL can verify the proof.
Call Evidentia from your backend, avoid sending raw AI data by default, and use public verification without exposing raw prompts or outputs.
Open Beta trust scope
Current signing infrastructure
Current Open Beta uses platform-managed signing infrastructure intended for technical evaluation. It validates the verification flow, not production-grade key custody.
Proof integrity.
Consistency of proof metadata.
Tamper detection.
Reproducibility of verification.
Future production infrastructure
Production trust infrastructure is planned
Persistent production-grade key management, hardware-backed signing, external key attestation, and independently auditable signing infrastructure are planned for future production environments.
Persistent signing identities.
KMS/HSM-backed signing.
Externally auditable trust infrastructure.
Key rotation policies.
Independent attestation.
Steps
Integration flow
Create your workspace
API key generated
Store the API key in your backend environment, Secret Manager, or CI/CD secrets.
Send Proof requests from your backend to Evidentia.
Check proof status
Store proof_id, verification_url, public_proof_api_url, and status in your database.
Share a public verification URL only after a proof is intentionally published.
Security
API key handling
Evidentia API must be called from your backend server.
Never expose API keys in frontend code, mobile apps, GitHub, browser storage, or logs.
API keys belong only in server-side environments such as .env files, secret managers, or CI/CD secrets.
Data minimization
Send proof metadata, not raw data
Do not send raw prompts, raw outputs, personal data, or confidential business data unless you explicitly choose to.
During Open Beta, send hashes, summaries, and metadata whenever possible.
Idempotency
Use the header as the source of truth
The Idempotency-Key header is required. Use the same value for the same AI usage record to prevent duplicate proof requests.
Use the same Idempotency-Key for the same AI usage record to prevent duplicate proof requests.
Client logs
Keep your application logs
Operational debugging
Internal records
Analytics
Business workflow history
Evidentia Proof
Add verifiable proof
Tamper-evident proof
Third-party verification
Audit and dispute evidence
Proof that a decision happened
Reference the same decision_id in your client log and Evidentia Proof so internal records and external proof stay safely linked.
Client storage
Store the IDs you will need later
After a Proof is requested and completed, store these values in your own system.
decision_id
idempotency_key
job_id
proof_id
verification_url
public_proof_api_url
status
created_at
related_internal_log_id
Proof request
POST /api/proof/request
Request proof. Check proof status. Store proof_id and verification_url.
Share public verification URLs with auditors, customers, or third parties only for proofs that have been intentionally published. During Open Beta, customer proofs are private by default.
Evidentia does not replace your application logs. Send a payload that references, summarizes, or hashes the decision, then store the returned proof identifiers in your database.
Example code
Start from the repository examples
Use the Node.js / Express example as a base, then adapt it to your backend.
Fixed demo and system proofs use Solana devnet anchoring. Mainnet, multi-chain verification, and dedicated infrastructure are planned production capabilities and are not self-service during Open Beta.
Current Open Beta uses platform-managed signing infrastructure intended for technical evaluation. Developer access is self-serve, paid plans are request-only, and live billing remains gated.
Verification export packages are available for public proofs and authorized private proofs; raw AI payloads are not included by default.