Skip to content

Documentation

Plans & limits

What the free plan includes, what counts against your budget, and how paid plans are provisioned.

Free-plan defaults

At signup, free-plan organizations start with:

  • 2 workspaces
  • 2 jurisdictions per workspace (must select explicitly; empty is not allowed)
  • 1 seat
  • 15 tracked bills
  • 10 tracked regulations
  • 3 language watches
  • 25 stakeholders
  • ~50k AI tokens / month for org-scoped generation (onboarding wizard assists, Ask Billy chat, draft alerts, and similar)
  • No developer API or outbound webhooks — those ship on standard / enterprise (see Developer API)

Org admins can see live usage versus limits in Settings → Organization (seats, workspaces, tracked bills, watches, AI tokens, reset period).

Live usage versus free-plan caps in Settings → Organization — AI tokens, workspaces, tracked bills, and language watches.

What counts as AI usage

  • Org-scoped: wizard profile suggestions, workspace chat, draft client / action alerts, newsletters, outreach drafts, and similar interactive generation — these draw down the monthly token budget.
  • Autonomous agentic: overnight triage, morning brief, and draft-ahead are platform-funded with separate daily caps — they do not burn your seat quota; when exhausted they defer. See Agentic assists.
  • Global pipeline: normalization, embeddings, and structural diffs are shared discovery substrate. LLM summarization for tracked/suggested bills is globally memoized (orgs share the same cached summary rows) and metered under a separate system budget — not billed as if every org recomputed every bill.

Higher workspace / seat / tracked-bill / watch caps, Slack, and broader digest options are sales-provisioned via platform admin — not a self-serve upgrade button in Phase 1. Contact the Billy team if you need more capacity for a design-partner or firm rollout.

Consistency guarantee

Returning to the same version pair shows byte-identical analysis until inputs change. Affirmed summaries remain the system of record for that review even if a later generation supersedes the AI text for new readers (your review stays attached to what you saw).