Head to head · Content management · October 2026 research run

Sanity vs Storyblok

Sanity scores 75.2 (BB) on agent readiness against Storyblok's 67 (B), and leads in 5 of 7 scored categories. Storyblok leads on security & auth. Both do content management.

Best headless CMS and website publishing for AI agents · All 136 cms comparisons

Which one, for what

Sanity BB

Best for Teams that model content as structured documents and want an agent to draft, patch and stage changes in releases for a person to publish.

Ahead on

  • Reliability, 87 against 64
  • Schema & documentation, 87 against 68
  • Payments & pricing, 60 against 53
  • Maintenance & community, 89 against 80
  • Transparency & trust, 87 against 81

Also in its favour

  • Agent-ready, a grade of BB or better

Watch for

Schema validation rules run in Sanity Studio and, since October 2026, in MCP writes. The HTTP mutation API accepts a document without checking them

Storyblok B

Best for Teams already on Storyblok who want an agent to draft, translate and publish stories or change component schemas under scoped, per-space permissions.

Ahead on

  • Security & auth, 73 against 67

Also in its favour

  • Free to start without a card

Watch for

The Management API's OpenAPI spec sits in a private repository. Only the Content Delivery API has a public spec (OpenAPI 3.1, 14 operations)

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Sanity +23

Sanity 87

Graded on the hosted HTTP API and the hosted MCP server. Statuspage at www.sanity-status.com with 25 components, including api.sanity.io, apicdn.sanity.io and MCP Server, and a full incident history (20). The 22 July 2026 incident on api.sanity.io, open for 6 hours 20 minutes, which held this line at 10 on 17 October, left the 90-day window on 21 October. Between 3 August and 1 November the incidents filed against api.sanity.io or apicdn.sanity.io are short or partial. On 14 August API and API CDN errors ran from 14:34 to 14:40 UTC (labelled major). On 27 October a small subset of customers saw intermittent latency and some errors from 13:37 UTC until a fix at 18:54, resolved at 21:27 (labelled major, "most customers are not affected"). Smaller ones fell on 6 August (errors and timeouts for a small subset, 09:25 to 09:45), 20 August (timeouts for a small subset, 13:05 to 13:45), 15 September (mutation errors), 14 October (latency for some customers), 20 October (524 errors for a small subset, 14:56 to 15:16) and 26 October (13:12 to 15:13, a small subset). Partial incidents read as minor in this re-check, however they are labelled. The 8 October critical and the 8 to 9 October major listener incidents (Studio, Functions and Dataset Embeddings), the 23 September Content Agent incident and the 28 October critical on Studio and Dashboard, whose update says the Content APIs were not affected, are not filed against the API or MCP components. No MCP Server incident in the window. Minor incidents only (20, from 10, cause time). Rate limits are published with numbers, 500 requests a second per IP, 25 mutations and 25 uploads a second per IP, 500 concurrent queries and 100 concurrent mutations per dataset, per the 7 October check (15). 429 is documented, @sanity/client retries queries with exponential backoff and the docs tell callers to queue mutations, and writes can be made safe with transactionId, ifRevisionID and createIfNotExists. No Retry-After header is documented for the HTTP API and there is no idempotency-key header (12). The SLA at sanity.io/legal/sla commits to 99.9 per cent on Enterprise E1 and 99.95 per cent on E2 (10). The HTTP API is generally available and the MCP server has been since v2.6.0 on 11 December 2025 (10). 20 + 20 + 15 + 12 + 10 + 10 = 87.

Storyblok 64

Graded on the Management API and the hosted MCP server. Public UptimeRobot status page at uptime.storyblok.com, awarded 15 of 20 on 7 October for four delivery monitors and none for the Management API or the MCP server. Its monitors and history load by script from /api/ paths that the status host's robots.txt disallows, so on 1 November 2026 we couldn't re-read which parts of the service it covers, and the 7 October reading of 90 days at 100 per cent came from the same disallowed path and no longer counts (15 of 20, as awarded on 7 October). The incident record is unread, so the line takes the no-readable-history score, a limit of our reading and not a finding against Storyblok (5). Rate limits published, 3 requests a second on Starter and 6 on paid plans (15). The docs advise slowing down on 429 and retrying with exponential backoff, and the official client retries on 429. No Retry-After header in the API docs and no idempotency keys found (9). SLA of 97 per cent on Growth and Growth Plus, 99.9 on Premium and 99.99 on Elite, stated for the Content Delivery API (10). Management API v1 is generally available and the MCP docs carry no beta label (10). 15 + 5 + 15 + 9 + 10 + 10 = 64.

Schema & documentation Sanity +19

Sanity 87

26 OpenAPI specs covering 225 operations are public at www.sanity.io/docs/api/openapi and through sanity openapi get (25). llms.txt, llms-full.txt and a Markdown copy of every documentation page (10). All 225 operations carry a description, and the MCP tool descriptions say when to use a tool and when not to, though several are one line (16). The specs hold 702 enums, but documents, patches and query results are free-form JSON by design and a query is one GROQ string (10). 1,272 example keys across the specs. 141 of 225 operations document a 4xx response and 9 document 429, and the Mutation reference lists 400, 401, 403 and 404 without response bodies (11). Dated API versions in the URL and a public changelog with an entry per release (15).

Storyblok 68

The only public OpenAPI spec is for the Content Delivery API (3.1, 14 operations). The Management API spec that the official client is generated from sits in the private storyblok/openapi-wdx repository, per the monoblok README. The MCP describe tool returns parameter and body schemas per operation to a signed-in client, which we couldn't read (12 of 25). A root llms.txt and a Markdown copy of every docs page at the same URL plus .md (10). The MCP page says when to use the server and when to use the CLI, and names each tool's purpose (16). The reference types every parameter and marks required ones, but story content is a free-form object checked against the space's component schemas (8). Request examples in nine languages on each endpoint. Errors are covered by one line on standard HTTP codes (9). The /v1 path, a dated changelog with an RSS feed, and changes to v1 made in place (13).

Maintenance & community Sanity +9

Sanity 89

MCP server v2.45.2 on 29 October 2026 and Sanity Studio v6.21.0 on 28 October 2026 (30). The MCP registry lists 34 versions of the server published between 3 August and 29 October 2026, and the changelog has an entry for each recent one (20). The open-source Studio repository had 6,352 stars and 284 open issues and pull requests. Two of the three newest open issues had comments, a small sample, per the 7 October check, not re-read on 1 November. There is a Discord channel for the MCP server (17 of 25). io.sanity.www/mcp is in the official MCP registry under Sanity's own domain at version 2.45.2, and @sanity/client is at 8.14.1 (15). Packages are current and require Node 22.12 or later. We didn't read their CI (7 of 10).

Storyblok 80

The changelog has an entry on 26 October 2026 and the Management API client 0.11.0 and CLI 4.26.0 were tagged on 30 October 2026 (30). 15 dated changelog entries and more than 100 package tags in the monoblok repository in the 90 days from 3 August (20). Public changelog with RSS, a help centre and a Discord community. Of the five newest GitHub issues on 7 October, two were closed within a week and three had no reply after 9 to 14 days, not re-read on 1 November because GitHub's pages refuse our reader (10 of 15). Current official SDKs and CLI, though no Storyblok entry in the official MCP registry, searched on 1 November (13). The monoblok repository has CI, release and commit-lint workflows and 44 commits from 7 October to 30 October. We didn't check that CI passes (7).

Payments & pricing Sanity +7

Sanity 60

No x402, MPP or L402 in the documentation or on the pricing page (0). Plans and unit prices are public. Growth is $15 a seat a month with overage at $1 per 25,000 API requests, $1 per 250,000 API CDN requests, $0.30 a GB of bandwidth and $0.50 a GB of assets (20). The Free plan runs without a card, with 10,000 documents and 250,000 API requests a month (20). A person has to create the account in a browser. After that, projects, datasets and robot tokens can be created through the API or the MCP server (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 40 is 60.

Storyblok 53

No x402, MPP or L402 (0). Plan prices are public ($99 and $349 a month) with unit prices for overage on Growth ($10 per million API requests, $15 a seat, $20 a locale). Premium and Elite are priced through sales (15). Starter is free with no card and includes the Management API (20). A person signs up in a browser and creates the token or approves the OAuth consent screen (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 35 is 53.

Security & auth Storyblok +6

Sanity 67

The MCP server uses OAuth with PKCE, dynamic client registration and a revocation endpoint, but lists one scope, global. The HTTP API takes robot tokens that carry a role, can be deleted at any time and since June 2026 can be given an expiry, with a written rotation procedure (24 of 30). A Viewer token is read-only on every plan, MCP edits land in drafts or release versions and publishing is a separate tool, and the separate Context MCP server is read-only. No confirmation step is documented, and roles limited to a dataset or document type are Enterprise only (15 of 20). No prompt-injection guidance found in the MCP documentation. Content comes from project members and not the public, and run_sanity_cli runs without a shell or filesystem access (3 of 15). Document history records each transaction, the Activity Log API covers management events, and request logs can be exported for 7 days. The activity feed is absent on Free and the full audit trail and History API are listed under Enterprise (11 of 15). SOC 2 Type 2, a vulnerability disclosure programme and a security.txt without an Expires field. The bug bounty pilot has closed and no ISO 27001 certificate of Sanity's own was found (14 of 20).

Storyblok 73

OAuth with PKCE S256, dynamic client registration, refresh tokens, revocation and 29 scopes on a read, write and publish hierarchy, or personal access tokens with scopes, a space list and an expiry date (30). Less 5 because Content Delivery API tokens travel in the token query parameter. That is a judgement call. Those tokens are read-only and not the graded surface, so we took half the checklist's 10. Scopes separate read, write and publish per space, and execute_destructive needs explicit user confirmation. Unscoped legacy tokens work until 7 December 2026, moved from 30 November (18). The changelog names prompt-injected tokens as a risk that scoping limits, and the MCP page asks users to read plans back before writes. No guidance on treating stored content as untrusted (8). Activity log on every plan with an Activities endpoint, kept 1 day on Starter and 30 days on Growth (11). ISO 27001, TISAX, third-party penetration tests and a security contact address. No security.txt, bug bounty or SOC 2 report found (11).

Transparency & trust Sanity +6

Sanity 87

Sanity Studio, the CLI and @sanity/client are MIT. The Content Lake and the MCP server are closed services under published terms (20 of 30). The privacy policy of 4 May 2026 says logs are deleted or anonymised within 90 days and document history is kept for a period set by the plan, which the pricing page gives as 3 days on Free, 90 on Growth and 365 on Enterprise. The DPA of 12 August 2026 covers return and deletion, and links the sub-processor list at an older URL (25 of 30). API versions due for removal are announced by email and on the website, marked with X-Sanity-Deprecated and answer 410 once removed, and deprecated APIs are labelled in the reference. The notice period is described as appropriate, with no number of days (14 of 20). The sub-processor list of 19 January 2026 (unchanged on 1 November) names 11 companies with processing locations, with Google Cloud in Belgium as the primary host (20).

Storyblok 81

Closed service with published self-service and enterprise terms, and MIT client libraries (17). DPA effective 22 April 2026, privacy policy effective 6 October 2025, customer content kept up to 90 days after termination, log files 30 days, and AI terms saying input isn't used for training. We found no contradiction between them (25). No written deprecation policy, but dated notices, such as about six months for unscoped tokens (7 December 2026, moved on 23 October from 30 November) and about a month for Management API validation (13). Sub-processors with locations in DPA Annex 3 and data residency in the EU, US, Canada and Australia (18).

Score by category

CategoryWeightSanityStoryblokEdge
Reliability16%8764Sanity +23
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%8768Sanity +19
Agent ergonomics13%7478Storyblok +4
Security & auth14%6773Storyblok +6
Payments & pricing10%6053Sanity +7
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8980Sanity +9
Transparency & trust7%8781Sanity +6
Negative events≤15-3-3
TotalBB 75.2/100B 67/100

Facts side by side

FactSanityStoryblok
KindHTTP APIHTTP API
VendorSanity US Inc. and Sanity ASStoryblok GmbH
Hosted endpointhttps://api.sanity.iohttps://mapi.storyblok.com/v1
TransportsHTTP, Streamable HTTPHTTP, Streamable HTTP
AuthOAuth or keyOAuth or key
PricingFreemiumFreemium
x402nono
LicenceProprietary hosted service under Sanity's terms of service. Sanity Studio, the CLI, @sanity/client and the agent toolkit on GitHub are MITProprietary service under Storyblok's terms. The SDKs, API clients and CLI in storyblok/monoblok are MIT
Tools exposed587
Read-only variant documentednono
llms.txtyesyes
MCP registryio.sanity.www/mcpnot listed
Last release2026-10-292026-10-30
Terms last updatedno date givencouldn't be read
Privacy policy last updatedno date given2025-10-06
Customer content may train modelsnot found in the textcouldn't be read
Terms restrict automated accessnot found in the textcouldn't be read
Terms restrict benchmarkingyescouldn't be read
Terms or service can change without noticenot found in the textcouldn't be read
Arbitration or class-action waivernot found in the textcouldn't be read
Popularity6.4k stars, 4.1M npm/wk68 stars, 433k npm/wk
Agent reviews3.3/5 (4)3.3/5 (4)

The panel on both

The 3 reviewers who reviewed both rate them the same.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

How the scores have moved

Verdicts

Sanity

Sanity publishes 26 OpenAPI specs covering 225 operations, and its hosted MCP server saves edits to drafts or release versions, with publishing as a separate call. The Content Lake does not run schema validation on HTTP writes, custom roles limited to one dataset or document type are Enterprise only, and MCP tools have changed with notice only on release day.

Storyblok

The hosted MCP server covers the whole Management API with seven tools, OAuth scopes split into read, write and publish per space, and a confirmation step on deletes. The Management API has no public OpenAPI spec and no idempotency keys, its status history couldn't be read, and publishing is a GET request.

Before you call either

Sanity

  1. Pin a static dated version in every URL, such as v2025-02-19. Omitting apiVersion in @sanity/client falls back to v1.
  2. Validate documents against the schema yourself before an HTTP write, or run sanity documents validate afterwards. The Content Lake does not enforce schema rules.
  3. Back off on 429 for mutations yourself. @sanity/client retries queries five times but never retries mutations. The limit is 25 mutations a second per IP.
  4. Over MCP, call create_version before patch_documents when editing inside a release, then patch the returned version ID with the same releaseId.
  5. Use GROQ projections and slices to size results. MCP query responses are limited to 64 KiB, and a blocked Free project answers 402 with plan_limit_reached.
  6. Re-read the MCP tool list when a call fails with an unknown tool. Sanity has removed and renamed tools on release day

Storyblok

  1. Pick the base URL by the space's region (mapi.storyblok.com for the EU, api-us, api-ca or api-ap otherwise). A token sent to the wrong region fails
  2. Create stories without publish to keep them as drafts, then call the publish endpoint. It is a GET, so never prefetch or blindly retry it
  3. Write translations as field__i18n__<code> keys inside the same content object, and set the component field to translatable first
  4. Upload an asset in three steps (signed response, POST to S3, finish upload). Through MCP the S3 step needs shell access for curl
  5. Stay under 3 calls a second on Starter and 6 on paid plans, and back off exponentially on 429. Saves that break a field's max_length return 422
  6. Replace unscoped personal access tokens with scoped ones before 7 December 2026, when Storyblok revokes them

Questions

Which is better for AI agents, Sanity or Storyblok?

Sanity scores 75.2 (BB) on agent readiness against Storyblok's 67 (B), and leads in 5 of 7 scored categories. Storyblok leads on security & auth.

Do Sanity and Storyblok need an API key?

Both take an API key or an OAuth sign-in.

Can an agent call Sanity and Storyblok without installing anything?

Yes. Sanity has a hosted endpoint at https://api.sanity.io and Storyblok at https://mapi.storyblok.com/v1.

Other comparisons with Sanity or Storyblok

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.