Skip to main content

Evidence Export Pack

Evidence Export is the feature teams use when governance stops being a local engineering concern and starts needing to be shared with other people.

That usually means one of these situations:

  • security wants a review pack instead of API access
  • compliance wants evidence that governance controls are running
  • a platform team needs a snapshot of governed activity before an upgrade or production rollout
  • leadership wants a concrete record of what AI workflows are actually doing

Instead of forcing reviewers to query the system directly, AxonFlow can bundle the relevant governance records into a single structured export.

What The Export Contains​

Depending on the requested record types, an evidence pack can include:

  • audit_logs
  • workflow_steps
  • hitl_approvals

That makes it useful across both standard governed request traffic and workflow-oriented systems that include approval gates.

Endpoint​

POST /api/v1/evidence/export

Example Request​

curl -X POST "http://localhost:8080/api/v1/evidence/export" \
  -H "Content-Type: application/json" \
  -d '{
    "start_date": "2026-02-15",
    "end_date": "2026-03-01",
    "types": ["audit_logs", "workflow_steps", "hitl_approvals"],
    "limit": 1000
  }'

Request Fields​

FieldRequiredDescription
start_dateYesStart of the export window, using YYYY-MM-DD or RFC3339
end_dateNoEnd of the export window, defaults to now
typesNoRecord types to include; if omitted, all supported types are included
limitNoMaximum records to return, capped by tier

Response Shape​

{
  "export_id": "3a7c9e2f-1b4d-4e8a-9c6f-2d5e8b1a3c7f",
  "tenant_id": "my-org",
  "tier": "evaluation",
  "date_range": {
    "start": "2026-02-15T00:00:00Z",
    "end": "2026-03-01T00:00:00Z"
  },
  "disclaimer": "NOT FOR REGULATORY SUBMISSION - EVALUATION LICENSE",
  "record_count": 142,
  "audit_logs": [],
  "workflow_steps": [],
  "hitl_approvals": [],
  "exported_at": "2026-03-01T12:00:00Z",
  "daily_usage": {
    "used": 1,
    "limit": 3
  }
}

Response Fields​

FieldMeaning
export_idunique identifier for the export
tenant_idtenant the data belongs to
tiercurrent tier at export time
date_rangeeffective export window
disclaimerevaluation watermark, omitted in enterprise
record_counttotal records returned
audit_logsaudit log rows included in the pack
workflow_stepsworkflow-step evidence included in the pack
hitl_approvalsapproval records included in the pack
exported_atexport timestamp
daily_usagequota information where the tier has limits

Evidence Summary​

If you want to see what is available before exporting the full pack, use the summary endpoint:

GET /api/v1/evidence/summary
curl -X GET "http://localhost:8080/api/v1/evidence/summary" \

That endpoint returns counts over the tier’s evidence window, which is often the fastest way to sanity-check whether the records you care about are actually present.

Tier Behavior​

CapabilityCommunityEvaluationEnterprise
Evidence Export❌✅✅
Evidence Summary❌✅✅
Max lookback window❌14 daysUnlimited
Max exports per day❌3Unlimited
Max records per export❌5,000Unlimited
Watermark / disclaimer❌Evaluation disclaimerOmitted
Scheduled and broader operational reporting❌❌✅

Evaluation is intentionally useful for pilots and internal reviews, but not a substitute for the long-term operating surface enterprises usually need.

Why This Matters For Production Readiness​

Evidence export becomes valuable surprisingly early. Teams often think they will only need it for formal audits, but in practice it is also useful for:

  • pre-launch readiness reviews
  • internal architecture reviews
  • incident postmortems
  • comparing governance behavior before and after a policy change

It is one of the clearest examples of why Evaluation can be a strong bridge tier. You can build with Community, but once people outside the immediate engineering team need evidence, a shareable export surface becomes hard to replace with ad hoc scripts.

Typical Review Workflows​

What makes this feature especially valuable is that it is not tied to one regulator or one review style. The same export mechanism can support:

  • internal platform readiness reviews
  • security investigations
  • privacy or compliance evidence gathering
  • executive or architecture reviews that need concrete artifacts

Security Review​

Export a recent window of audit logs and workflow activity for a security or privacy review:

curl -X POST "http://localhost:8080/api/v1/evidence/export" \
  -H "Content-Type: application/json" \
  -d '{
    "start_date": "2026-02-22",
    "end_date": "2026-03-01",
    "types": ["audit_logs", "workflow_steps"]
  }'

Approval Review​

Focus on human oversight decisions:

curl -X POST "http://localhost:8080/api/v1/evidence/export" \
  -H "Content-Type: application/json" \
  -d '{
    "start_date": "2026-02-15",
    "end_date": "2026-03-01",
    "types": ["hitl_approvals", "workflow_steps"]
  }'

Upgrade Baseline​

Use GET /api/v1/evidence/summary to inspect what is available, then export a baseline pack before changing policies, infrastructure, or license tier.

What Reviewers Usually Want Included​

A useful evidence pack usually tells a coherent story, not just a large one. In practice, reviewers often want:

  • the relevant time window
  • audit logs tied to the affected tenant or workflow
  • workflow-step records for governed execution
  • approval records where human review mattered

That is why the types filter matters. It lets teams export a reviewable slice instead of dumping every available record.

Why This Matters Commercially Too​

Evidence export is one of the strongest examples of how the docs should guide readers from technical use to commercial fit:

  • Community is enough to build and understand the governance model.
  • Evaluation is where teams can start sharing bounded evidence with security and compliance stakeholders.
  • Enterprise is what larger organizations usually want when evidence handling becomes a recurring operating function rather than a one-off export.
Last updated on