Pricing
Start free. 10K device checks a month, no card.
One device check is your app asking us to confirm a request came from real hardware. Pro is $99 for 250K checks. Business is $899 for 2.5M. Overage gets cheaper as you scale. Move the calculator below to find your cheapest plan.
Plans
One clear comparison
Free
$0
10,000 device checks / month
- Works with every TPM chip vendor out of the box
- All built-in strictness policies
- Dashboard, live event feed, per-account ban list
- Community support
Pro
$99 / month
250,000 included · $1.50 / 1K overage
- Everything in Free
- Customer-authored acceptance policies
- Email support · < 24h response
Business
$899 / month
2,500,000 included · $0.75 / 1K overage
- Everything in Pro
- Uptime + latency SLA
- Priority email + Slack channel support
Enterprise
Custom
Annual commit · volume pricing
- Everything in Business
- SOC 2 audit packet (in progress)
- Old API versions keep working 90 days after we change them
- Single-tenant deployment available
- Named CSM · on-call escalation path
Calculator
Find your cheapest plan
Slide the inputs to your real numbers. The calculator picks the cheapest tier that covers your volume, shows the bill math, and estimates what you save versus the fake signups you absorb today (industry baseline: about $2 for every 1,000 that slip past a CAPTCHA).
From 100 to 1,000,000 (log scale)
Includes step-up + re-attest cadence (default 1.2)
% of attempted signups that are currently bots / fakes
Effective attestations / month
60,000
50,000 signups × 1.2 att/user
Your monthly Root Herald bill
$99.00
$99 base — 60,000 attestations ≤ 250,000 included
Recommended tier
Most popularPro
Pro covers your volume at the lowest total cost.
Projected savings vs CAPTCHA solver farms
—
At your current abuse rate, solver-farm cost is below the platform fee. The win is structural — once fakes can't bypass, attackers stop trying.
Feature matrix
Every feature, every tier
| Feature | Free | Pro | Business | Enterprise |
|---|---|---|---|---|
| Volume | ||||
| Included device checks / month | 10K | 250K | 2.5M | Custom |
| Overage price (per 1K) | — | $1.50 | $0.75 | Volume |
| Hard quota cutoff | ||||
| Platform | ||||
| Works with every TPM chip vendor | ||||
| Built-in strictness policies | ||||
| Write your own strictness policies | ||||
| Deployment | ||||
| Default (rootherald.io) | ||||
| Single-tenant deployment | ||||
| Support & SLA | ||||
| Community / docs | ||||
| Email support | < 24h | Priority | Named CSM | |
| Shared Slack channel | ||||
| Uptime + latency SLA | ||||
| SOC 2 audit packet (in progress) | ||||
| Old API versions keep working after a change | 90 days | 90 days | 90 days | 90 days |
Comparison
What you'd otherwise pay
Different tools, different failure modes. Here's the honest read: cost, user friction, and what eventually defeats each.
| Approach | Cost | User friction | Defeated by |
|---|---|---|---|
| CAPTCHA solver farms (Turnstile / hCaptcha bypass) | $1–$3 per 1,000 solves to attacker | Visible challenge to real users | Defeated by solver farms within hours |
| Fingerprinting (FingerprintJS Pro) + anti-detect browsers | $200/mo for 100K + $30–60/mo anti-detect license | Silent · but blocks legitimate privacy-conscious users | Defeated by GoLogin, Multilogin, AdsPower, Kameleo + residential proxies |
| KYC (Persona, Sumsub, Onfido) | $0.50 – $3 per check + PII handling cost | Government ID + selfie: kills signup conversion | Doesn't bind to device; valid ID + valid selfie still abusable at scale |
| Root Herald Shield | $0 for 10K · $99 for 250K · $899 for 2.5M | Silent · invisible to users · no PII | Cheap attacks (bots, cloud VMs, emulators) blocked outright; serious attacks now cost ~$30 of real hardware per device — and show up as a detectable cluster |
FAQ
Pricing FAQ
What counts as a device check?
One metered unit per attestation check that reaches appraisal — the point where we actually evaluate the device evidence against a policy. That includes signup checks, step-up re-attests, and periodic posture re-checks. The eventual verdict doesn't change the bill: a pass, a warning, and a contraindicated (failed) verdict each count as one check, because each one did the work.
Requests rejected before appraisal are not billed: an invalid API key (401), an unknown policy (422), a malformed request (400), or an over-quota request (429) never reaches the verifier, so it never meters.
Do retries count?
Each challenge nonce is single-use, so retrying the same nonce can't be billed twice: a consumed or expired nonce is rejected rather than re-metered. Retries with a fresh nonce (e.g., the user got logged out and re-attested) count separately, which matches reality, since each is a separate device check with a separate cryptographic verdict.
What happens when I hit my Free-tier quota?
The Free tier has a hard quota cutoff. Once you hit 10,000 device checks in a calendar month, the API returns HTTP 429 quota_exceeded, and the quota resets at the start of the next calendar month. Paid tiers do not cut off; they bill overage at the per-1K rate listed.
You can upgrade in-app at any moment. The change applies immediately; the calendar-month quota resets and overage billing kicks in from that moment forward.
What if I need a self-hosted / on-prem deployment?
Enterprise tier only. We'll ship you a single-tenant container stack that stays current with the latest chip-maker trust roots and keeps your token-signing keys safe. Pricing is on a per-deployment basis; talk to sales.