On-device hashing
Chunk hashes and Merkle roots cover the captured bytes to make tampering detectable.
Epistemic Recorder
Every recording includes a signed proof you can verify locally (Verifier CLI) or at ingestion (Epistemic Vault). Recorder computes hashes and signs a manifest during capture, with randomized capture-time challenges recorded in the proof.
Tamper-evident recording for verifiable media: cryptographically verifiable video, proof-of-origin media, and forensic-grade recording that preserves authenticity from capture to verification.
Proudly EU-based.
Available as Android and iOS apps, plus the Epistemic Universal Box (EUB) for external camera workflows. EUB is a touchscreen recorder (Atomos Ninja-style) that generates the same proofs as the apps and can connect directly to the Vault.
Epistemic Recorder on GitHub Verifier CLI on GitHubRecorder signals
Integrity and provenance are created at the source, not added later.
Lightweight capture UX with heavy-weight guarantees.
Chunk hashes and Merkle roots cover the captured bytes to make tampering detectable.
Canonical CBOR payloads wrapped in COSE signatures for portable verification.
During capture, Recorder issues randomized challenges that are recorded in the proof and visible in verification output.
Attestation + host integrity signals summarize capture conditions (e.g., device integrity).
Same proof model, different form factors.
Mobile
Capture in the field with device-backed keys and embedded provenance.
Hardware
A dedicated touchscreen recorder for external cameras—same proofs as the apps.
Workflow
Upload to Vault with proofs intact for zero-trust ingestion and receipts.
Designed to move with your team without breaking flow.
Record on Android/iOS, or via EUB with external cameras and pro monitoring workflows.
Manifests are signed and embedded inline (no fragile sidecars required).
Connect directly to Vault (or any verifier) to independently recompute and validate proofs.
Trust layer
Recorder is built so Vault can re-verify every byte at ingestion and issue a signed receipt.
This does not try to “prove reality” from pixels alone. It verifies the capture proof and whether the file matches it.
Verification works wherever the file goes.
Trust conditions are encoded, not implied.
Preserve provenance across edits and distribution.