100 USERS

Evidence API · v1

From product evidence to the next decision.

Read critic records, trace findings to source receipts, advance fixes, run matched retests, and record outcomes without confusing modeled signals for verified results.

Explore endpoints
Public readsVerified-owner writesSigned service intakeEvidence-class boundaries

Evidence

Read the aggregate report

Public

Return product history, findings, critic opinions, interventions, and linked outcome receipts.

GET/api/v1/products/:slug/report
Request contract
  • slug: product URL identifier
Response contract
  • product
  • versions[]
  • findings[].opinions[]
  • findings[].outcomes[]
  • disclosure
curl 'https://100-users.com/api/v1/products/nike/report'

Example response

{"data":{"product":{"slug":"nike"},"versions":[],"findings":[],"disclosure":"Synthetic and modeled evidence is separated from verified human and commercial outcomes."}}

Output model

A report is a connected evidence graph.

Every aggregate remains inspectable down to the version, session, critic opinion, intervention, retest, and outcome that produced it.

01

Product + version

Artifact identity, observed version, methodology, and comparison lineage.

02

Evidence receipt

Source, content hash, visibility, capture time, actor, and session.

03

Finding + opinion

A bounded issue plus critic interpretation, confidence, and dissent.

04

Intervention

Proposed fix, priority, workflow state, and linked evidence.

05

Matched retest

Comparable cohort, task, methodology, and measurement window.

06

Outcome

Before and after values, trust state, source receipt, and milestone eligibility.

07

Critic record

Stable constitution, evidence history, prediction, calibration, and disclosure.

08

Human calibration

Participant counts, comparisons, agreement, status, and boundary.

Outcome integrity

Trust is a field, not a footnote.

Only source-verified and independently reviewed outcomes count toward verified commercial milestones.

modeled

A projection or simulation, never observed revenue.

submitted

A customer-supplied result with a retained source receipt.

source_verified

The metric source was verified and counts toward the commercial milestone.

independently_reviewed

The source and interpretation received independent review.

Verified-owner workflow

Claim, fix, retest, prove.

  1. 1

    Request a domain-matched email challenge.

  2. 2

    Verify the code and receive a 24-hour owner token.

  3. 3

    Advance the evidence-backed intervention.

  4. 4

    Queue a matched retest with an Idempotency-Key.

  5. 5

    Record an outcome with its trust state and source receipt.

Signed service authentication

Service integrations sign timestamp + "." + rawBody with HMAC-SHA256 and send the result as x-innovation-signature. Never expose the shared secret in browser code.

Operational contract

Predictable writes and explicit failures.

Error envelope
{"error":{"code":"validation_error","message":"Explain what failed","details":[]}}
Status codes

400 invalid input · 401 authentication · 403 ownership · 404 unknown record · 409 conflict · 429 rate limit · 500 internal failure

Idempotency

Retests, outcomes, and service applications require an Idempotency-Key. Reusing it returns the existing receipt instead of creating a duplicate.

Planned, not callable in current v1

Test creation, live event streams, and outbound webhooks.

The current API exposes persisted critics, evidence, interventions, claims, matched retests, outcomes, calibration, waitlist intake, and signed applications. Direct POST /tests, test event streams, and webhook subscriptions remain roadmap contracts, not production routes.