Guide · CompanionGuard · Published 14 November 2026 · Reviewed 14 November 2026 · By the Ingenious AI team
CompanionGuard Plans Compared: Free, Pro at $29 and Team at $59
The short answer
Free covers one bot's user-visible kit: disclosure snippet, hosted protocol page, crisis-resource component, minor break reminders and referral counts. Pro ($29/month) adds the evidence layer: the eval harness with 100 runs a month, the Crisis Evidence Runtime, the referral dashboard with annual-report CSV export, and the California and Colorado disclosure packs. Team ($59/month) scales it: 1,000 eval runs a month and multiple bots. The scenario-test runner is free for everyone.
What separates the CompanionGuard tiers?
The tiers separate visibility from evidence. Free makes the safety layer visible to users (banner, page, resources, reminders). Pro makes it provable (scored reply tests, sealed live decisions, packs, exports). Team makes it plural (many bots, ten times the eval runs). Every tier states the same boundary: a compliance toolkit, not legal advice, not a clinical service.
Create the free safety kit Next guide All guides
CompanionGuard is built by Ingenious AI. A compliance toolkit — not legal advice and not a clinical service. If you are in crisis right now: US — call or text 988; UK & Ireland — Samaritans 116 123.
The split, as the pricing page states it
| Free | Pro — $29/mo | Team — $59/mo | |
|---|---|---|---|
| Bots | 1 | 1 (packs + evidence) | Multiple |
| Disclosure snippet + hosted protocol page | Yes | Yes | Yes |
| Crisis-resource component + locale packs of resources | Yes | Yes | Yes |
| Minor break reminders | Yes | Yes | Yes |
| Referral counts | Yes | Yes | Yes |
| Scenario tests (12, free runner) | Yes | Yes | Yes |
| Eval harness | — | 100 runs/mo | 1,000 runs/mo |
| Crisis Evidence Runtime | — | Yes | Yes |
| Referral dashboard + annual-report CSV | — | Yes | Yes |
| CA + CO disclosure packs | — | Yes | Yes |
| Audit log (evidence chain) | — | Yes | Yes |
Prices and limits are the product's own: PLANS in the code (pro 29, team 59) and PLAN_LIMITS (100 and 1,000 eval runs per month). The free tier's completeness is deliberate — a safety surface that only paying bots show would leave most users of most bots unprotected, so the visible layer carries no price.
Who should stay free, and for how long
Stay free while your open questions are user-visible ones: Is the banner showing? Is the protocol page correct? Do the 12 scenarios pass? A pre-launch bot, a small community bot with no partner or platform scrutiny, and any builder in their first month should exhaust the free tier's discipline before paying for evidence of a discipline they have not yet built. The free tier also starts the counting clock — referral counts accumulate from creation — which a later Pro upgrade inherits, so 'free for now' costs your future records nothing.
The honest limit of free: you can say what your bot shows, but not prove how it behaves at scale or produce sealed records of live decisions. When a conversation starts needing the word 'prove', the tier conversation starts too.
The Pro triggers
Four triggers recur. A reviewer asks: an app store, a partner or a school wants evidence, not intentions — the harness summary and evidence export are the answer's shape. A jurisdiction conversation opens: California or Colorado duties become concrete for your user base, and the packs plus annual-report CSV move from nice to necessary. An incident happens: after the first real crisis event, 'we tested and here is the sealed record of decisions' is the only retrospective sentence worth having. Your change cadence rises: weekly prompt or model changes need the harness baseline as a release gate, and 100 runs a month is sized for exactly that rhythm on one bot.
Any one trigger justifies Pro on its own arithmetic: $29/month against the cost of one unanswerable diligence process or one unevidenced incident. The pricing page carries the current offer, including the manual PayPal door and crypto checkout the fleet supports.
When Team is the actual shape of your problem
Team is not 'Pro but prouder'; it answers two specific shapes. Multiple bots: several characters, several products, or a white-label fleet — one safety operation across all of them, with per-bot evidence. Run volume: 1,000 eval runs a month serves teams testing many bots after every change, or one bot with an aggressive experimentation cadence. If neither shape is yours, Team buys you headroom you will not use, and the honest advice of this guide is to stay on Pro until a shape changes.
The multi-bot operational pattern — shared baselines, per-bot chains, fleet-level counts — is covered in the multi-bot guide. And the refund frame for any paid step is the product's stated clause: once a license key is activated the sale is final; an unactivated key that is not working is fixed or refunded via [email protected].
Five actions to choose the right tier
- Exhaust the free discipline first: banner verified, protocol page correct, 12 scenarios passing, counts accumulating.
- Name your trigger for Pro in writing — reviewer, jurisdiction, incident-readiness or change cadence — so the upgrade is a decision, not a drift.
- Estimate your eval rhythm honestly (changes per month × bots × re-runs) against 100 runs before assuming you need 1,000.
- Count your bots: more than one production bot is Team's defining shape, not a detail.
- Revisit the tier quarterly; audiences and scrutiny change faster than pricing does.
Worked example: three bots, three tiers, one year
A fictional builder's year in tiers: January–March, a single pre-launch companion on Free — kit embedded, scenarios passing, counts accumulating from the first test traffic. April: a school partnership asks for evidence; Pro ($29/month) is switched on the same week, the harness baseline is set at 93, the runtime starts sealing decisions, and the partnership signs with the protocol page and harness summary as its safety exhibits.
November: two sibling characters launch for the holiday season. The builder moves to Team ($59/month) — three bots under one safety operation, 1,000 monthly runs absorbing a test-after-every-change rhythm across all three, per-bot evidence chains feeding one quarterly export habit. Total spend across the year stays under a single hour of legal consulting, which the builder notes in the retrospective with some feeling. The story is invented; the prices, limits and feature split are the product's own.
Frequently asked questions
What does the free tier include?
One bot's complete user-visible kit: the disclosure snippet, hosted safety-protocol page, crisis-resource component with locale packs (988 US + others), minor break reminders and referral counts. The 12-scenario test runner is also free.
What does Pro add for $29/month?
The evidence layer: the eval harness (100 runs/month), the Crisis Evidence Runtime with hash-chained audit log, the referral metrics dashboard with annual-report CSV export, and the California and Colorado disclosure packs.
What does Team add for $59/month?
Scale: multiple bots and 1,000 eval runs per month, with the full Pro feature set across the fleet.
Are the disclosure packs really paid, if disclosure is a duty?
The disclosure mechanisms (snippet, protocol page, crisis line) are free. The packs — the jurisdiction-specific duty lists and banner variants for California and Colorado, applied at creation — are a Pro inclusion. Builders remain responsible for their own legal position either way; the product is a toolkit, not legal advice.
What are the refund terms?
Per the product's stated clause: digital delivery — once a license key is activated, the sale is final; if a key has not been activated and is not working, contact [email protected] and it will be fixed or refunded.
Keep going
Related guides: Eval Run Budgeting: 100 a Month on Pro, 1,000 on Team — Who Needs What · Multiple Bots, One Safety Operation: Running CompanionGuard on Team · Indie Character Bot Safety: Big-League Guardrails on a Solo Budget
Put it to work: Create the free kit · Eval harness · Scenario tests · Crisis runtime · Disclosure packs · Pricing · All CompanionGuard guides