Seekapedia: API data sets

Plans

Verify, Monitor, Govern

Three jobs a compliance team does with license data: check a business, keep it under observation, and manage the organization's response. Every plan gets the same trustworthy data: source links, freshness, unresolved conflicts, and a clear line between what the agency reports and what your policy decides.

Pricing is set per plan when you start a pilot. Everything below is what the API does today.

Verify

Check a business during onboarding or review.

Available today

  • Look up a license by identifier, or match a business by name and address.
  • A match outcome you can act on: exact, probable, candidates to review, or no match.
  • Source link, freshness and conflict information on every answer.
  • Bulk submissions, up to 500 per request.
  • Saved, downloadable check results as evidence.
  • Your own policy applied to the result, kept separate from the data.
  • Test webhook events.

Monitor

Detect changes across a portfolio you already work with.

Available today

  • Everything in Verify.
  • Saved portfolios with an exception queue: suspended, past-expiration, missing from the latest source, unresolved conflicts, overdue data.
  • Signed webhook alerts when a watched license changes, with retries, delivery history and replay.
  • The change feed and a per-license change history.
  • Versioned saved policies.
  • A usage report that counts monitored locations.
  • Email alerts to addresses the owner has confirmed: a digest of changes after each update, plus review notices.
  • A read-only dashboard: plan and usage, data freshness, open reviews, portfolio exceptions, deliveries and the audit log.

Govern

Manage reviews, policies and evidence across teams.

Available today

  • Everything in Monitor.
  • Immutable policy versions; every check records the version it was judged against.
  • Team keys with four roles: viewer, reviewer, approver and admin.
  • Review assignments, escalation, and approvals where the person who submits a decision cannot approve it.
  • Decision packages: the data as it was when the review opened, the data now, and every step taken.
  • A tamper-evident audit log with exports and a one-call chain check.
  • Single sign-on through your identity provider, with OpenID Connect or SAML 2.0, and roles mapped from your groups.
  • Automatic user and group provisioning and deactivation from your identity provider (SCIM 2.0).
  • Historical policy evaluation: judge a license, or a whole portfolio, as our data stood at an earlier time, under the policy version in force then or any other; and re-judge a saved check or review under a different policy.

Not available yet

  • History from before our first sync (we hold history from 19 November 2026 onward).
  • Contracted response commitments.

Offered through a scoped proposal: the review, audit, sign-on and provisioning features above are built and tested against simulated providers; we set each one up with you against your real identity provider. Response commitments are not available.

How plans and limits work

Each account is on one plan: Verify, Monitor or Govern. The plan decides which endpoints the account can use; a request for something the plan does not include is refused with 403 plan_does_not_include and says which plan has it. Limits on monthly checks, monthly lookups and monitored locations are set per account when we agree them, and GET /v1/usage shows your plan, your limits, what you have used and what remains. Over a monthly limit, the request that crosses the line completes and the next one gets 429 quota_exceeded with the time the count resets. Accounts without a plan assigned have everything and no limits.

How we count

Billing units are defined before prices so you know what you are buying. The same definitions are returned by GET /v1/usage, which reports your current counts.

Standard check
Verify
One completed POST /v1/checks for one submitted business location. A result of no match or candidates is a completed check. Not counted: polling or reading a saved check, repeating a request with the same Idempotency-Key, requests we reject as invalid, and our own technical failures.
Monitored location
Monitor
One canonical premises per customer account, however many licenses, watches or portfolios it appears in. A store with three licenses counts once. A license with no usable address counts on its own. The count is measured now: removing a watch takes it out at once, archiving a portfolio does not remove its watches, and unresolved candidates are not monitored until you add a specific license. How a part-month is billed is a contract term, not a counter.
Platform scope
Govern
Set by proposal: platform commitment, included portfolio size and volume bands. Policy rules and webhooks are not priced per item, so you are never discouraged from using them.

Four things that are easy to confuse

DimensionWhat it means here
Historical data coverageAgency records are current-state. Our change history starts at our first sync (November 19, 2026) and grows daily. It does not reach back into an agency's own history.
Your evidence retentionSaved checks stay until you delete them. Longer retention does not mean we hold a longer agency history.
Source refreshWe read each source once a day at 14:45 UTC. Each source's own publication schedule is listed at /v1/sources.
Notification latencyA notification after our system detects a supported source change (daily sync, then hourly delivery and retries). It is not an immediate alert when a license's legal status changes: agency publication timing is outside our control.

Enforcement and conditions

Not available yet. Today's data covers license status and identity only. Every answer says so (enforcement_review_state: "not_covered"), because the absence of enforcement data is not evidence that none exists.

We have looked for published, machine-readable enforcement and license-condition data from the Texas, New York and California agencies and have not found any we can use, so this is not on a delivery date.

When enforcement and conditions data is added, it will be a coverage-based module that ships with its coverage schedule (jurisdictions, agencies, document types, history start dates, refresh targets, known omissions). Allegations will be kept separate from findings, and stayed or conditional penalties will always be shown with the penalty itself, never held back as an upgrade.

Pilots

Intended planA pilot should demonstrate
VerifyRepresentative submissions, defensible matches, useful evidence, and how exceptions are handled.
MonitorPortfolio setup, relevant change delivery, replay, and reconciliation. Synthetic test events can exercise delivery and failure handling when no live change occurs.
GovernWho reviews and who approves, how exceptions are routed, and whether a decision package and the audit export give your auditors what they ask for. We test sign-on and provisioning with your identity provider early, since each needs a registration on your side.

We agree success criteria before starting. "We found no adverse records" is not evidence of accuracy, because this data does not cover enforcement.

Request a pilot