Pricing reference
Glyphward pricing — free, $29, $99
Three plans. One rate card. No “contact sales,” no seat multipliers, no per-region surcharge. Here is what each plan includes, how much scanning that gets you, and when it makes sense to move up a tier.
TL;DR
Free is 10 scans a day for trying, testing, or hobby use. Pro at $29/mo is 100,000 scans a month — a real production budget for an indie AI app. Team at $99/mo is 1,000,000 scans, SSO-lite, and Slack alerts for companies with more than one engineer touching the integration. Every plan covers image and audio scanning; no modality is paywalled.
The full rate card
| Plan | Price | Included scans | Rate limit | Built for |
|---|---|---|---|---|
| Free | $0 / mo | 10 / day (~300 / mo) | 2 req/sec | Hobbyists, researchers, benchmark runs |
| Pro | $29 / mo | 100,000 / mo | 20 req/sec | Indie AI apps shipping image or voice |
| Team | $99 / mo | 1,000,000 / mo | 100 req/sec | Funded startups with shared ops |
What is a “scan”?
One scan is one image or one audio clip sent to the API. A user upload scored for FigStep, AgentTypo, typographic PI, and audio-waveform anomaly counts as one scan, not four — we bill by call, not by detector. Clips up to 30 seconds and images up to 8 MB are in-bounds without surcharge. Longer clips and larger images still scan; they just consume one quota unit each.
Working out which plan you need
A quick usage calculation: if the user-facing feature is “upload a photo, get a response,” expect roughly one scan per upload per day-active user. Scale from there.
- Under 300 scans/month: Free covers it. This is hobby volume or a single-digit-DAU staging environment.
- 300 to ~100k scans/month: Pro. An avatar SaaS with ~500 DAU doing 5 upload-triggered scans a day comfortably fits under the cap.
- Over 100k scans/month: Team. Also the right plan if you need SSO-lite (one-tenant SAML, no IdP gymnastics) or the Slack alert channel for new-attack-vector notifications.
Overages and rate limits
Free and Pro both soft-cap rather than hard-fail. If you cross the included quota on Pro, subsequent scans are billed at $0.00035 per scan (3.5 tenths of a cent) up to 2x the plan, at which point we email and ask you to confirm an upgrade. No silent overage bills. Free-tier overage returns an HTTP 429 with a clear message — you don’t get charged for spilling over.
Rate limits above are per API key. You can issue multiple keys (e.g. staging vs prod) at no extra cost on Pro and Team; the quota is shared across all keys on the account.
Why it is priced this way
The market anchor above us is Lakera Guard at $99 and up for the self-serve tier before moving upmarket post-Check Point acquisition — see Lakera alternative (multimodal) for the longer story. The market anchor below us is the open-source LLM Guard (free, text-only). Glyphward is deliberately priced to be a clear upgrade from “we have nothing for multimodal” without requiring a procurement cycle — the Pro plan covers most indie-AI-app production use for less than what most teams already spend on a single transactional-email sender. A full comparison of the tier is on the pricing comparison page.
Related questions
Is there an annual discount?
Not at launch. Once we have real revenue data we’ll offer a 2-month-free annual on Pro and Team — current plan is to ship it in the first 90 days post-launch. Founding Pro subscribers lock their rate for 12 months.
Is there a per-call price without a subscription?
Not at launch. The Free tier exists precisely to cover one-off and benchmark usage without a card. A usage-based no-subscription tier is on the backlog for customers who fit neither 10/day nor 100k/mo — it needs Stripe metered billing, which is coming but not first.
Can I get a refund?
Yes, pro-rated, no questions, within the first 30 days. After that the plan is cancellable at any time and runs to the end of the current billing month.
Is audio scanning at the same price as image scanning?
Yes — one scan is one scan, regardless of modality. Audio is not upsold separately. This is a deliberate anti-pattern break: most “multimodal” tooling paywalls the second modality. We don’t.
What happens to my existing rate when you raise prices?
Existing subscribers keep their rate for 12 months from signup. Price changes apply only to new sign-ups during that window. If you care about locking a rate, the Free tier also serves as a grandfathered-slot reservation — sign up once, upgrade later, keep the original rate.
Further reading
- Multimodal prompt-injection scanner pricing comparison (2026) — how the tiers line up against Lakera, LLM Guard, Azure Prompt Shields, and Promptfoo.
- Lakera alternative for multimodal prompt injection — why self-serve tooling under $100/mo matters for our ICP.
- FigStep detection — one of the concrete attacks every plan covers.