Plans and usage
What each plan includes, how protected agents and prompt volume are counted, and how usage warnings work.
Plans
| Plan | Protected agents | Prompts / month | Seats | Retention |
|---|---|---|---|---|
| Free (evaluation sandbox) | 0 | 5,000 | 1 | 7 days |
| Production Pilot (30 days) | 3 | 1M | 10 | 90 days |
| Launch | 10 | 10M | 10 | 90 days |
| Scale | 40 | 50M | 25 | 365 days |
| Enterprise | 150, more by contract | 100M, or contracted volume | Contracted | 730 days |
- Free: gateway sandbox, Red Team Lab (limited), basic findings. Not a commercial protection tier.
- Production Pilot: everything in Launch at pilot volume, for 30 days before an annual agreement, with an end-of-pilot report.
- Launch: gateway, runtime detections, Tool Guard and the MCP gateway; human approvals in Slack, Teams and email, and plan approvals; runtime authorize tools (inspector, approval metrics, ask-user); PII and secrets detection; findings and evidence exports; Guardian and HIRO; attack-surface mapping; a standard SOC 2 evidence pack.
- Scale: everything in Launch plus SIEM export, custom policy packs, scheduled red teams, HIRO investigations on a stronger model, 1-year retention.
- Enterprise: everything in Scale plus customer-hosted deployment (pilot), advanced compliance mapping (NIST AI RMF, GDPR, EU AI Act) and executive risk reporting.
The approvals inbox (list, read, approve, deny, vote) is on every plan, because Intertrace can hold a call on every plan. Your plan and usage are in Settings → Billing.
Protected agents
Protected agents are counted across the organization. Every way of protecting an agent checks the same quota; at the limit it refuses with HTTP 402 and Protected agent limit reached (used/limit). Remove protection from another agent or upgrade your plan.
Prompt volume and warnings
Prompt volume is the number of governed requests this UTC calendar month, the same number Settings → Billing shows. Authorized security tests and rate-limited requests are not counted.