Sanity pricing

About

Sanity is a headless CMS that positions itself as a “Content Operating System.” Its two pillars are the Content Lake — a hosted, real-time content database queryable through GROQ and GraphQL APIs — and Sanity Studio, an open-source, React-based editing environment that teams customize to match their workflows. Content is modeled as structured data, validated once and reused across websites, apps, and other channels, rather than locked into page templates. More recently Sanity has layered AI on top of this foundation (Content Agent, Agent Actions, and AI Assist), all metered through a shared AI-credits pool.

Sanity serves the full range from individual developers on its Free plan through global brands on Enterprise. Customers named across its pricing and enterprise pages include PUMA, Riot Games, AT&T, SKIMS, Mejuri, Morning Brew, and Tecovas. It competes with Contentful, Storyblok, and legacy stacks such as WordPress and Adobe Experience Manager, differentiating on a developer-first, content-as-data model, an open-source editor, and — as of 2026 — deeply integrated, usage-metered AI tooling.

Pricing summary : How Sanity’s seat + pay-as-you-go usage model works

Sanity uses a hybrid seat-plus-usage model with three dimensions:

1. Per-seat subscription: Free ($0, up to 20 seats) and Growth ($15 per seat/month, up to 50 seats); Enterprise is custom-quoted with custom seat counts.
2. Pay-as-you-go usage: Growth includes monthly quotas (25K documents, 1M API CDN requests, 250K API requests, 100GB assets, 100GB bandwidth) and meters overages — $1 per 250K CDN requests, $1 per 25K API requests, $0.50 per 1GB assets, $0.30 per 1GB bandwidth, $299 for 50K documents, and $999 per additional dataset.
3. AI credits: every organization gets a monthly pool (1,000 credits on Free and Growth, 5,000 on Enterprise); additional credits are $0.05 each, with a Content Agent query costing 4 credits and agent actions 1–2 credits.

What makes this different: the base seat price is deliberately low ($15), so most of the bill scales with metered project usage and AI consumption rather than with seat count.

Pricing by product

Sanity platform (self-serve plans)

Sanity platform (Enterprise)

Growth plan add-ons

Pay-as-you-go usage overages (Growth)

Sales motions across products: PLG / self-serve for Free and Growth (card-billed); sales-led for Enterprise (custom-quoted, invoiced).

AI credits (how Sanity meters AI)

Sanity’s AI tooling — Content Agent, Agent Actions, and certain MCP-server tools — is metered in AI credits at $0.05 per credit. A query (your message to Content Agent) costs 4 credits ($0.20); an action or tool use costs 2 credits ($0.10); a request that triggers an Agent Action costs 1 credit ($0.05). Every organization receives a free monthly pool — 1,000 credits on Free and Growth, 5,000 on Enterprise — and can set a monthly AI spending limit that pauses all AI operations once reached (an in-progress operation runs to completion). Directional examples from Sanity’s docs: “Update 5 documents” ≈ 34 credits, “Translate 3 documents into 2 languages” ≈ 76 credits.

Hidden costs : What Sanity teams actually pay once usage overages stack up

The “$15/seat” headline understates what busy content teams actually pay. Because Growth meters almost every dimension separately, the overages and AI credits often rival or exceed the seat subscription. Two representative examples:

A 10-seat marketing team with a high-traffic site

Here the seat subscription is only ~14% of the bill — bandwidth and the support add-on dominate, which is invisible if you budget from the sticker price alone.

A 6-seat team leaning heavily on Content Agent

At ~12,000 credits (roughly 3,000 Content Agent queries at 4 credits each), AI consumption alone is over six times the seat cost — a reminder that agent-metered features can quietly become the largest line item.

Want to estimate your own Sanity bill? Use the Sanity pricing calculator to model your monthly cost based on seats, usage overages, and AI credits.

Pricing evolution : How Sanity moved toward multi-dimensional usage metering

Cadence

Tracked range: 2020–2026. Quarters not listed above were verified stable (0 price changes, 0 SKU additions).

Notable changes

• 2020 — Sanity priced per project: a pay-as-you-go Standard tier (from $0) and an Advanced tier (from $199/project/month), each metering extra users, datasets, API requests, assets, bandwidth and documents.
• 2022-01 — Project tiers renamed Free / Team / Business / Enterprise; Team started at $99/project/month, Business at $949/project/month, with extra non-admin users at $15/user.
• 2023-H2 — The defining shift: Sanity dropped per-project pricing for a per-seat Growth plan at $15/user/month, cut asset ($0.50/GB) and bandwidth ($0.30/GB) overage rates, and launched four Growth add-ons.
• 2025-H1 — Functions invocations, compute GB-seconds, Agent Actions and Media Library storage were added to the metered pricing table.
• 2026-02 — An explicit AI Credits meter appeared: 1,000 free credits/month on Free and Growth, 5,000 on Enterprise, $0.05 per additional credit; Content Agent queries cost 4 credits and actions 2 credits.
• 2026-06 — The $1,399/month SAML SSO Growth add-on was removed; SSO with SAML moved to Enterprise-only, corroborated by third-party pricing coverage.

The 2023 model overhaul in detail

Sanity’s most consequential pricing decision was retiring per-project pricing. Through 2022, a customer with several projects paid a project-level subscription (e.g. Team at $99/project/month) plus per-project overages — a structure that penalized teams running many small sites and made costs hard to reason about across a portfolio. The 2023 rebuild collapsed this into one organization-level model: a low $15/seat Growth base, generous shared quotas, and pay-as-you-go metering on the dimensions that actually track cost (requests, bandwidth, storage, documents). It simultaneously lowered the marginal overage rates (bandwidth fell from $1/5GB to $0.30/GB) while adding higher-priced convenience add-ons, shifting Sanity from “pay for projects” to “pay a little for access, then pay for what you consume.”

What’s unique : Distinctive mechanics in Sanity’s usage-metered CMS pricing

1. A deliberately low seat price with the real money in metering. Where most headless-CMS peers anchor on a seat or platform fee (Contentful’s paid tier starts near $300/month), Sanity’s Growth seat is just $15. That makes the seat-plus-usage model function almost like pure usage-based pricing: the seat is an access token and the bill grows with API requests, bandwidth, documents and AI credits rather than headcount.

2. Almost every dimension is a separate meter. Sanity bills seats, API CDN requests, API requests, assets, bandwidth, documents, datasets, function invocations, compute GB-seconds, embeddings reads and AI credits as independent line items — an unusually granular metered per-request design. It gives large customers fine-grained cost attribution but also means no single number predicts the invoice.

3. AI is credit-metered, not bundled. Content Agent and Agent Actions draw from a shared AI-credit pool priced at $0.05/credit (a query is 4 credits, an action 2). Rather than raising the seat price to “include AI,” Sanity kept the base flat and put AI on its own credit-based meter with a free monthly allotment — so AI-heavy teams self-select into higher spend.

4. Datasets are a premium, discretely-billed unit. Beyond the two included datasets, each additional dataset is a hefty $999/month add-on (max four on Growth). Treating the dataset — Sanity’s unit of content isolation — as a high-priced SKU is a distinctive lever that pushes multi-brand and multi-environment teams toward Enterprise.

5. Overage rates fall as the model matures. Unusually, Sanity has repeatedly cut marginal usage rates over time (bandwidth from $1/5GB in 2021 to $0.30/GB by 2023), even while adding higher-margin add-ons — a sign it treats raw usage as a growth-lubricant and monetizes convenience and scale instead.

Strengths & weaknesses

Billing UX : Manage-console controls, AI spending limits, and usage metering

• Add-ons in Manage — enable Growth add-ons (Dedicated Support, Increased quota, Extra datasets) from Manage → select project → the Plan tab → Add-ons; each add-on is billed as a line item on the next billing cycle’s invoice.
• AI spending limits — set a monthly cap under sanity.io/manage → Settings; when the organization hits the cap, AI features pause until the next billing cycle or until the limit is raised (an operation already running finishes even if it exceeds the remaining budget).
• Usage tab — sanity.io/manage → Usage shows AI consumption over time and surfaces which individuals are the most prolific AI users.
• Pay-as-you-go metering — Growth projects automatically meter usage beyond the included quotas at the published per-unit rates ($0.30/GB bandwidth, $0.50/GB assets, $1/250K CDN requests, etc.), added to the monthly invoice.
• Spending-confirmation prompts — for large bulk AI operations, Content Agent may request confirmation before proceeding if estimated credit usage is high.
• Billing method by tier — Free and Growth pay by card; Enterprise gets invoicing and organization-level billing. Annual payment is available on request (per the pricing-page FAQ), but self-serve cards are billed monthly in USD.

Strategic wins : Why Sanity’s low-seat, usage-metered decisions worked

1. Anchoring on a low seat price, not a platform fee

By pricing the Growth seat at $15 instead of a flat platform subscription, Sanity made the entry decision trivially easy and pushed monetization into consumption. This is a textbook value-metric decision: the seat is the low-friction “get in the door” price, and the API/bandwidth/AI meters capture value from the customers who actually scale. It also neutralizes the “too expensive for a small team” objection that peers with $300 floors invite.

2. Keeping AI on its own meter instead of raising the base

Sanity resisted the temptation to bundle AI into a higher seat price. Content Agent and Agent Actions are metered in credits at $0.05 each with a free monthly pool, so AI-curious teams can try it without a price hike and heavy users pay proportionally. This mirrors the broader move toward outcome- and usage-aligned AI pricing and avoids taxing the majority who don’t yet use AI.

3. Cutting overage rates as the model matured

Rather than milking early usage, Sanity repeatedly lowered marginal rates (bandwidth from $1/5GB to $0.30/GB) while introducing higher-value add-ons. Treating raw usage as cheap growth fuel and monetizing convenience (support, quota bumps, extra datasets) is a disciplined take on usage-based packaging that keeps the meter from becoming a growth tax.

4. Fully public, self-serve pricing right up to Enterprise

Every non-Enterprise number — seats, quotas, overage rates, add-on prices — is published, so buyers can model costs without a sales call. Transparency at this depth is a competitive weapon in a category where rivals gate details, and it feeds AI-search and comparison tools that Sanity then wins on price for small teams.

Areas to improve : Specific gaps in Sanity’s pricing with proposed fixes

1. The bill is hard to predict across a dozen meters

Sanity meters seats, four request/traffic dimensions, storage, documents, datasets, functions, compute and AI credits separately — powerful for attribution, punishing for forecasting. A prospective buyer cannot answer “what will this cost me” from the pricing page alone. Fix: ship an official interactive cost estimator (as some third parties already have) and a “typical bill” example per archetype, the way usage-based leaders communicate unpredictable cost.

2. Cliff-edge add-ons distort the smooth-metering story

Most meters scale gently, but the $999/dataset and $799 support add-ons jump the bill in large discrete steps that clash with the pay-as-you-go narrative. Fix: offer a metered or tiered dataset price (e.g. per-dataset-day) and a smaller support tier so mid-market teams aren’t forced from $0 to $799 for a single question.

3. AI credit burn is opaque at the point of use

A Content Agent query silently costs 4 credits and multi-document actions can run to dozens; teams only discover the burn rate after the fact in the Usage tab. Fix: surface a live per-action credit estimate in the Studio before an operation runs, plus projected month-end AI spend, so editors can self-regulate the way good FinOps guardrails intend.

4. Removing self-serve SSO pushes security-conscious teams to sales

Withdrawing SAML SSO as a Growth add-on means any team needing SSO must jump to custom-quoted Enterprise. Fix: keep a self-serve SSO add-on (even at a higher price) or fold basic SSO into Growth; security should not be the single lever that forces a sales motion for otherwise self-serve customers.

Key takeaways

1. A low seat price can be a wedge, not the business model. Sanity’s $15 seat exists to remove the entry decision; the revenue lives in metered API, bandwidth, document and AI consumption. If your value scales with usage, price the seat to get in the door and monetize the meter.
2. Meter AI separately with a free pool. Putting Content Agent on a $0.05 credit meter with 1,000 free credits lets non-users avoid subsidizing AI and lets heavy users pay for what they burn — a cleaner path than a blanket seat-price increase.
3. You can cut overage rates and still grow revenue. Sanity lowered bandwidth and asset rates over time while adding higher-value add-ons; cheap raw usage drives adoption, and convenience/scale SKUs capture margin.
4. Granularity is a double-edged sword. A dozen independent meters give enterprises great cost attribution but make self-serve buyers anxious — pair granular metering with clear estimators and archetype examples.
5. Migrating pricing models resets your comparables. Sanity’s 2023 jump from per-project to per-seat means historical cost comparisons no longer hold; plan for how a model change breaks existing customers’ mental math and budgets.

UBP implications

1. Seat-plus-usage can approximate pure usage pricing. When the seat is priced near-nominal, the model behaves like consumption pricing with a light access fee — a pragmatic bridge for buyers uncomfortable with a fully variable bill while still aligning revenue to usage.
2. AI credits are becoming the standard AI value metric. Sanity metering Content Agent in credits ($0.05 each, 4 per query) echoes a corpus-wide pattern of decoupling AI cost from seats; expect credit pools with free allotments to be the default way software prices agentic features.
3. Transparent, granular metering favors comparison-driven buyers. Publishing every meter and add-on price makes a vendor legible to AI-search and cost tools — winning price-sensitive small teams — but raises the bar for predictability tooling so the granularity doesn’t scare mid-market buyers away.

Sources

• Sanity pricing page (accessed 2026-08-08)
• Sanity enterprise page (accessed 2026-08-08)
• Growth plan add-ons docs (accessed 2026-08-08)
• How AI credits work (docs) (accessed 2026-08-08)
• Sanity changelog (accessed 2026-08-08)
• Sanity blog (accessed 2026-08-08)

Bottom line

Sanity’s pricing tells a clear story: charge almost nothing for the seat, then meter everything else. A $15 Growth seat gets a team in the door, but the real bill is assembled from API requests, bandwidth, documents, datasets and — increasingly — AI credits at $0.05 apiece. That granularity makes Sanity cheap and transparent for small teams and precisely attributable for large ones, while quietly shifting the cost center to usage and AI. The trade-off is predictability: the more you grow, the more meters you have to watch.

Want to compare Sanity against other headless-CMS and usage-metered pricing? Browse the pricing blueprint.