Cryptographic provenance · Content-addressed storage

Verify any
data point independently

Every response from the AssayChain API carries a cryptographic attestation UID and a content-addressed CID. No central authority required — verify the source data directly from the public attestation registry and content-addressed storage.

Two parties, no trusted middleman. The seller cannot quietly change a delivered number; the buyer does not have to trust the seller's database — or ours.

What an attestation confirms: source, authenticity, and that the data was normalized on a given date. It does not certify that geological or assay values are true — closing that gap, and measuring how reliable the evidence is, is our core research (see Security).

There is no standard provenance layer for geological data

Legacy geological and assay data lacks a standardized integrity layer. Provenance is fragmented across PDFs, lab certificates, and agency databases, and there is no cryptographic link between a raw source document and the structured data extracted from it. A buyer who receives a number has no independent way to confirm it came from the document it claims to.

This is a trust gap with national consequence: critical-mineral supply-chain decisions, ESG disclosures, and reserve estimates are routinely made on data whose origin cannot be verified.

AssayChain is developing a provenance protocol that resolves it: a content-addressed attestation pipeline that creates a tamper-evident audit trail from source document to API response, without requiring trust in a central authority. The four-step pipeline below is that protocol in operation.

From source document to API in four steps

01

Source capture

A source document — a lab certificate, drill log, assay report, or agency record — is ingested and captured against a locked schema (assay-extract-v0.2): site, county, state, country, assayer, sample dates, and per-sample assays. Site coordinates are rounded to ±0.1° and personal identifiers are redacted before anything is stored. Field data can additionally be signed at the point of collection via an offline-capable capture prototype — a Phase I research objective.

02

Content-addressed storage

The completed record is stored in content-addressed storage — a network where each file gets a unique address derived from its content (called a CID). If the data ever changes, the CID changes too, making tampering immediately detectable. The file is content-addressed, not location-addressed: no central server can alter what the address points to. Attested source content is pinned for at least 5 years with a fallback host, so a verified record remains fetchable.

03

Verifiable attestation

A signed attestation is created encoding the record's data fields and content-addressed CID against a registered schema. The resulting UID is a permanent, independently verifiable record. The attesting key identifies the operator; key rotation and revocation are on the Phase I roadmap, so identity is governed rather than assumed permanent.

04

API delivery

When you call any paid endpoint via the metered API, the response includes attestation_uid and ipfs_cid. You can verify either independently, any time, without AssayChain's involvement.

Four registered schemas

All four schemas are registered and UID-committed in a public attestation schema registry.

Benchmark schema

Used for USGS commodity summaries (gold, silver, copper).

Schema name: mineral-intel-benchmark-v1

0xde9e029944dd16f3dad09458ec50b1823e4fbeaa35e6779509aeb84059bb8531

View on EASScan ↗

District report schema Live

Used for curated historical mining district reports (/historical). Revocable — a new attestation is issued when district data is enriched.

Schema name: mineral-intel-district-v1

0x4a7843b41982a42f574c2afab7bab592c0bcce1930ce3d3afae300a1084bf620

Committed fields: string district_slug, string source, string source_url, string result_cid, string attestation_date

View on EASScan ↗

Signed field capture Live

Signed-at-capture field data (offline-capable, device-keyed) — the Phase I prototype for attesting records at the point of collection, before they transit a server.

Schema name: mineral-intel-processing-run-v1

0x043a1d38402b3026ab8512a38d72fc0b88afbf3b6a95468e889212a582d2943a

Full measurements live in the content-addressed JSON referenced by the attestation's run_cid.

Committed fields: string run_id, string date_utc, string site_district, string feed_type, string device, string device_description, uint32 recovery_pct_estimated, string run_cid, string notes

View on EASScan ↗

Three ways to verify

Verify on EASScan

Take the eas_uid from any attested response. Go to base.easscan.org, search the UID. The full attestation — schema, data fields, attesting key, timestamp — is permanent and independently verifiable.

Verify content address

Take the ipfs_cid from any response. Fetch it via any IPFS gateway: https://ipfs.io/ipfs/<cid>. The returned JSON should match the data in the API response exactly — content-addressed, immutable.

Query EAS directly

Use the EAS SDK or GraphQL API to query all attestations against the run log schema UID. You can reconstruct the full attestation history without using the AssayChain API at all.

How the stack is composed

Layer Technology Role
Verifiable attestations attestation service (EAS) Permanent, typed records. Four schemas — USGS commodity benchmarks, district reports, price snapshots, and signed field capture (processing-run-v1).
Content-addressed storage IPFS Full record JSON stored before attestation. Each file is content-addressed — the CID is committed to the attestation, making tampering detectable.
Payment settlement HTTP 402 + settlement facilitator Core data free (rate-limited); paid calls from $0.10 USDC settled via a signed micropayment authorization. No accounts, no API keys.
API Cloud-hosted REST API ~30 endpoints across six categories — extraction, benchmarks, district history, price snapshots, and assistive intelligence. All data endpoints gated by the metered-payment middleware.
Web presence CDN hosting Dual-domain global edge delivery for resilience and availability. Agents that cache a base URL keep working through any single-domain incident — no retry logic, no failover code required. Fewer redundant HTTP retries means lower AI inference energy per successful call.
Signed field capture MineFlowTrack (PWA) MineFlowTrack — AssayChain’s offline-capable signed-capture prototype (Phase I) that attests field records at the point of collection.