We don't ask you to
trust our logs.
We let your auditor verify them.
Every action across every HASP surface — chat, documents, API calls, PHI scans, admin events — is signed with an Ed25519 key bound to your tenant, then chained. Any tampering breaks the chain at that point and everything after it. Your auditor downloads the export, runs the verification recipe on their own machine, and confirms integrity without any HASP software in the loop.
Every entry signed with a key bound to your tenant. Public key is published. Signature verifiable with standard cryptography tooling — no proprietary SDK.
Each entry references the hash of the previous one. You can't quietly edit a past entry — it breaks every entry that comes after it.
Chain checkpoints are countersigned by an external Time Stamping Authority. Timestamps don't depend on HASP's clock — they're attested by a third party.
Logs.
A dashboard with timestamps and event types. Written by the vendor, stored by the vendor, exported at the vendor's discretion. When an incident happens, the vendor tells you what the logs say. You have no independent way to confirm they haven't been altered.
- Stored and controlled by the vendor
- No way to detect retroactive modification
- "Trust us" is the audit answer
- Fails cross-examination in a contested audit
A signed, chained, independently verifiable record.
Every entry is cryptographically signed at write time. Entries are chained — each one references the hash of the one before it. Any modification to any past entry is detectable immediately, without asking HASP. Your auditor runs the verification on their machine; HASP is not in the loop.
- Signed at write time, chain is tamper-evident
- Tampering breaks the chain — detection is automatic
- Verification recipe runs on the auditor's machine
- Stands up under cross-examination
Coverage
Every surface. Every action. One chain.
The audit chain is not a separate product bolted on — it's the substrate every HASP surface writes to. There is no action you can take in HASP that doesn't produce a signed audit entry — including every PHI scan, where the categories detected and the action taken (allow, redact, or block) are recorded. For how detection differs from removal, see PHI scanning vs. redaction.
- Every chat message sent and received
- Every document upload and ingestion
- PHI scan: categories detected, action taken, user, prompt
- Model selection and token usage
- Every API call with caller identity (user / key / agent)
- PHI scan result per request
- Credit usage and budget enforcement events
- Rate limit enforcement and model allowlist enforcement
- Every Studio build conversation message
- Every app version created, previewed, and published
- Every app access event (authenticated user + timestamp)
- Rollback events with version restored
- Agent delegation issuance, refresh, and revocation
- Pre-action tool authorization allow / deny with rationale codes
- Agent Actions metering entries per approved invocation
- A2A handshakes binding external orchestrators to audit sequence IDs
- BAA countersignature and status changes
- Org member adds, removals, and role changes
- Audit export downloads (who exported, what range)
- Policy changes and model allowlist modifications
Verification
How your auditor verifies the chain
Six steps. Standard tooling. No HASP software required. Read the full verification recipe.
Download via the admin UI or the API. Output is plain JSON — one entry per line, with timestamp, actor, action, prev_hash, hash, and signature fields.
For each entry, compute SHA-256 of the canonical entry fields. Confirm it matches the stored hash. Confirm the stored hash matches the prev_hash of the next entry. Any mismatch identifies the break point.
Retrieve the tenant public key from /trust/keys/{org_id}. For each entry, verify the signature against the entry hash using the public key. Standard Ed25519 — works with OpenSSL, libsodium, Node crypto, Python nacl.
Periodic chain checkpoints include a TSA token. Verify the token using the TSA's public certificate (published). Confirms the chain existed before the timestamp — independent of HASP.
Filter the export for phi.scan.* events. Confirm every chat message has a corresponding scan entry. Confirm detected categories and the action taken (redact / allow / block) match your org's configured policy.
The recipe produces a machine-readable verification report: chain_valid, signature_valid, tsa_valid, phi_coverage, entry_count. Attach to your audit package. Done.
Capabilities
Everything in the audit chain
Chat turn, document upload, RAG retrieval, API call, BAA event, PHI detection, admin action — every audit-relevant event is signed with an Ed25519 key bound to your tenant. The public key is published.
Each entry references the cryptographic hash of the previous entry. Tampering with a past entry breaks the chain at that point and every entry after it. Detection is immediate.
Periodic chain checkpoints are countersigned by an external Time Stamping Authority. Timestamps don't depend on a HASP server clock — they're attested by an independent third party.
Any time period exportable as plain JSON plus a verification recipe. Auditors clone to their own machine, run the recipe, confirm the chain — without any HASP software in the loop. See the verification recipe.
Every PHI scan that fires gets one audit entry: which categories detected, which action taken (redact / allow / block), which user initiated, which prompt was scanned. Procurement asks; the answer is in the chain.
Seven-year retention on every tier. Monthly partitions for export performance. Chain remains continuous across partitions; verification covers the full retention window. Enterprise can extend beyond seven years on a custom contract.
Audit chain integrity is not a paid upgrade — signed, tamper-evident, and verifiable at every plan tier, including Solo.
Tier-1 wrappers say 'we're HIPAA compliant.' We show signed audit exports.
Reproducible verification on your auditor's machine — stands up under cross-examination.
Sample export and full verification recipe published — try it before you sign anything.
FAQ
The full platform
The audit chain covers all of these
Every HASP surface writes to the same chain. One BAA, one signed audit trail, one verification process — regardless of which surfaces your team uses.
A HIPAA-ready chat interface and document analysis tool for your whole team. Ask questions, get summaries, upload files — all with PHI scanning built in and every action on your audit trail.
Learn more →Describe the internal tool you need and watch it build live. No developers required — AI Studio generates a working app inside HASP, audited from the first keystroke.
Learn more →Add HIPAA-compliant AI to your own software. Drop in your existing AI SDK code, get PHI controls and a full audit trail enforced at the gateway — your integration stays clean.
Learn more →Connect external agents, automation pipelines, and A2A-protocol clients to HASP's policy gate. Every tool invocation is authorized, identity-scoped, and recorded to the signed audit chain — whether the caller is a human, a Studio app, or a fully autonomous agent.
Learn more → Try the verification
before you sign anything.
Download the sample export. Run the recipe. Confirm the chain holds on your own machine. Then decide if this is the audit story you want to bring to your next security review.