Content Marketing
Hygraph
EU-Made Report an error0–5 in half steps. 5 means the rubric's top anchor is met on the evidence.
by Hygraph GmbH · hygraph.com
Report an error on this page Is this your product? →
Link to this page
Add ?theme=dark to the image address for dark backgrounds.
The badge shows the category and the month of the latest evaluation, and EU-Made where
that is the product’s provenance. It never shows a score.
Hygraph at a glance
- Vendor
- Hygraph GmbH
- Headquarters
- Not stated on the vendor’s pages as of 13 Oct 2026
- Data hosting
- Not stated on the vendor’s pages as of 13 Oct 2026
- Free plan
- Yes, per the vendor’s own page source, captured 13 Oct 2026
- Panel verdict
- The AI panel rates Hygraph 2/5, the mean of 6 judges. Hygraph is a GraphQL-native headless CMS from Hygraph GmbH of Berlin. how scores are made
- As of
- Vendor pages captured 13 Oct 2026; evaluated by the panel 15 Oct 2026. all sources
Read this page as one judge. Each weighs the same scores by what they care about.
The panel's verdict
Hygraph is a GraphQL-native headless CMS from Hygraph GmbH of Berlin. Distribution and integrations draws the strongest scores, 6-7: webhooks are documented with HMAC-SHA256 signatures, a stated retry policy and logs, rate limits of 5, 25 and up to 500 uncached requests per second are published per plan, and a marketplace names connectors across commerce, assets, hosting, PIM and localization. Content model and reuse scores 4-7, with a genuine split: scores cluster at 6-7 on the schema editor, reference fields and five named translation connectors, while the Skeptic's 4 turns on portability — no documented export of content, model and locales. Weakest are editorial planning at 1-2 (content stages and comment assignment; no public information on calendars, briefs or workload views), AI and content rights at 1-3 (no public information on model, provider, prompt processing or training use), sovereignty at 0-3 (a Berlin contracting entity on the imprint; residency, ownership and subprocessor exposure recorded as unknown), and pricing transparency at 1-4 (concrete limits but no price figures, so an annual cost is not computable from the captured pages).
Speaks for it
- Webhooks documented end to end — HMAC-SHA256 signatures, a stated retry policy of 4 retries on top of the initial request, logs, pause and resume.
- Rate limits of 5, 25 and up to 500 uncached requests per second published per plan, alongside API token, webhook and content-stage limits.
- Schema editor with models, fields, components and enumerations, reference fields linking entries, and localized fields carried in webhook payloads.
- Marketplace names connectors across commerce, digital assets, hosting, localization and product information, including Lokalise, Crowdin, DeepL, EasyTranslate and Smartling.
- AI Assist shows every suggestion in a review panel before anything is written, and token usage is visible on the Billing page.
Held against it
- Planning on the captured pages stops at content stages and comment assignment; we found no public information on a calendar, briefs, campaigns or workload views.
- We found no public information on version history with restore, mandatory reviewers, guest or freelancer access, or an audit log of who published what.
- We found no public information on which AI model or provider processes prompts, where that processing happens, or whether customer content trains the models.
- Sovereignty attributes for data residency, ownership and subprocessor exposure are recorded as unknown, not confirmed on the vendor's own pages as captured.
- No price figures on the captured pricing pages, and no public information on seat costs, locale or record limits or overage rates, so an annual cost is not computable.
Best for
- You build composable sites and apps and need a GraphQL-native delivery API with signed, retrying, logged webhooks.
- You run multilingual content and want named translation connectors working alongside localized fields in webhook payloads.
- You plan editorial work in external tools and need structured, reusable content models with reference fields between entries.
Avoid if
- You need an editorial calendar, briefs, due dates and campaign grouping inside the CMS itself.
- Your procurement requires confirmed data residency, a published subprocessor list and a data processing agreement before signature.
- You must know which AI model processes your content, where, and whether it trains on it before switching AI features on.
- You need to compute an annual budget from publicly published prices before committing.
The scores
| Criterion | Panel median |
|---|---|
| Planning, editorial calendar & briefs | 1 of 5 |
| Creation, review & approval | 2 of 5 |
| Structured content, localisation & export | 3 of 5 |
| Publishing, distribution & integrations | 3.5 of 5 |
| AI assistance & content rights | 1 of 5 |
| European sovereignty | 1 of 5 |
| Pricing transparency | 1.5 of 5 |
Planning, editorial calendar & briefs
Show reasoningHide reasoning
How this is scored
Turning a content strategy into scheduled work: topics and campaigns, an editorial calendar across channels, briefs that carry the goal, audience and keywords to the writer, and a view of what is due, late or missing.
0 — No planning features; content is created as standalone entries with no calendar, status or owner.
3 — A shared calendar with publish dates and simple statuses, but no briefs, no campaigns or topics to group work under, and no notion of who owns a piece.
5 — An editorial calendar across channels with owners, due dates and custom statuses, briefs attached to each piece, and content grouped by campaign or topic cluster.
8 — Strategy is connected to the plan: topics or personas defined once and assigned to pieces, configurable workflows per content type, workload and deadline views per person, and recurring formats scheduled as series.
10 — The plan is accountable: every published piece traces back to its brief, topic and campaign, gaps in coverage per topic or funnel stage are reported, and plan-versus-published is visible over time without a spreadsheet.
The Head of Content
Comment assignment and comment statuses give me an owner and a state per entry, but that is where planning stops. We found no public information on an editorial calendar, due dates, briefs, campaign or topic grouping, workload views or recurring series — I cannot see what is late until it already is. 2 6 10
The SME Marketer
We found no public information on an editorial calendar, briefs, campaigns, topics or deadline views; the captured pages show content stages and the ability to assign a comment to a colleague, which is a lifecycle status rather than a plan. As the person who schedules our year by campaigns, there is nothing here to plan against. 2 6
The Localisation Lead
Content stages (Draft, Published and custom stages) and comment assignment give a basic notion of status and ownership, but we found no public information on an editorial calendar, briefs, campaigns or per-person deadline and workload views. Work is tracked as entries with stages, not as a scheduled plan. 2 6
The Web Engineer
Content stages cover Draft, Published and custom stages, and a comment can be assigned to a team member, but we found no public information on an editorial calendar, briefs, campaign or topic grouping, or per-person deadline and workload views. Structured entries with statuses exist; we found no public information on anything that turns a content strategy into scheduled work. 2 6 10
The Data Protection Officer
The captured documentation gives work a status and a responsible person — content stages from Draft and Published up to five custom stages on the top plan, and comments that can be assigned to a team member. We found no public information on an editorial calendar, briefs, campaigns or topic clusters, or due-date and workload views, so nothing on these pages connects a strategy to scheduled, accountable work. 2 6
The Skeptic
Draft and published stages plus the ability to assign a comment to a team member are the only planning-adjacent features the captured pages show. We found no public information on an editorial calendar, briefs that carry goal and audience to the writer, campaign or topic grouping, or per-person workload and deadline views. 2 6 10
Creation, review & approval
Show reasoningHide reasoning
How this is scored
Writing and reviewing content together: editing with comments and suggestions, roles and permissions for internal and external contributors, and an approval trail that records who released which version.
0 — Single-user editing with no comments, no roles beyond admin, and no record of changes.
3 — Comments and a version history exist, but every user can publish, external writers need a full seat, and approval is a status anyone can set.
5 — Roles that separate writing from publishing, inline comments and suggested edits, a version history with restore, and guest or freelancer access limited to assigned pieces.
8 — Configurable multi-step approval per content type or market, mandatory reviewers (legal, product, brand), scheduled publishing that waits for approval, and an audit log of every edit, approval and publish with user and timestamp.
10 — Approval is evidence: the exact version each approver released is preserved and exportable, permissions reach down to field and locale level, and the audit trail is retained for a stated period and available after the contract ends.
The Head of Content
Inline comments with mentions, nested replies and assignment, plus four system roles and admin-created custom roles, cover the review basics; a publishing workflows section appears in the plan comparison but with no captured detail. We found no public information on version history with restore, multi-step approval with mandatory reviewers, scheduled publishing that waits for sign-off, guest access limited to assigned pieces, or an audit log — nothing here yet says which version legal released. 2 6 7
The SME Marketer
Commenting is genuine collaboration — mentions, replies, resolve and assignment — and system roles separate creating, updating and deleting, with custom roles configurable by project administrators. We found no public information on guest or agency access below a full seat, on restoring earlier versions, or on an approval trail recording who released what; the pricing comparison hints at a publishing-workflows section without describing it. 2 6 7
The Localisation Lead
Inline comments with nested replies, mentions, resolution and assignment, plus four system roles and custom roles separating who edits from who administers, are documented in the editor guides, and webhook triggers include a workflow step transition. We found no public information on version history with restore, guest or freelancer access, mandatory reviewers, or an audit log of edits and approvals with user and timestamp. 6 7 10
The Web Engineer
Commenting is real — nested replies, @-mentions of people and roles, open/assigned/resolved states — and four system roles plus admin-created custom roles separate writing from administering. But we found no public information on entry version history with restore, guest or freelancer access limited to assigned pieces, configurable multi-step approval, or an audit log of who published what and when. The closest signals are custom content stages and a webhook Transition Step action. 6 7 10
The Data Protection Officer
Commenting with mentions, nested replies and open/resolved states, plus four system roles and custom roles built by project administrators, separates content work from administration, and a publishing workflows section appears in the plan comparison. We found no public information on version history with restore, guest or freelancer access limited to assigned pieces, mandatory reviewers, or an audit log recording who released which version. 2 6 7
The Skeptic
Commenting is real — nested replies, mentions of people and roles, resolve states — and roles separate creating content from administering the project, with custom roles built only by owners and project administrators. We found no public information on multi-step approval with mandatory reviewers, an audit log recording who released which version, guest or freelancer access limited to assigned pieces, or content version history with restore; publishing workflows surface only as rows in a plan comparison, though webhook triggers do include a workflow step transition. 2 6 7 10
Structured content, localisation & export
panel disagrees
Show reasoningHide reasoning
How this is scored
Whether content is stored as reusable structure rather than pages: a content model, components shared across channels, German and other locales managed side by side, and the whole repository exportable in a documented format.
0 — Content stored as formatted pages with no structure, one language per entry, and no export beyond copy and paste.
3 — Custom fields or templates exist, but localisation means duplicating entries, assets live elsewhere, and the only export is HTML or a CSV of titles.
5 — A configurable content model with reusable components, field-level localisation with fallback languages, a media library with alt text, and an export of entries and assets in JSON or a similar documented format.
8 — References between entries, variants per market or audience, translation workflows with status per locale and connectors to translation services named, and a full export of content, model, locales and assets documented and self-service.
10 — The repository is portable by design: model and content exportable together and re-importable, migration tooling or a CLI documented in both directions, version history included in the export, and no feature that only works while the content stays.
The Head of Content
The schema editor gives me models, fields, components, enumerations and reference fields, localized fields travel through webhooks under a localizations key, and five named translation connectors exist alongside a per-locale glossary. We found no public information on fallback languages, alt text in media handling, market or audience variants, or a documented export or migration tool for the whole repository. 1 9 10 11
The SME Marketer
The schema editor with models, fields, components and enumerations, plus reference fields linking entries, is real reusable structure, and named connectors to Lokalise, Crowdin, DeepL, EasyTranslate and Smartling cover our second-language workflow. We found no public information on field-level fallback languages, translation status per locale, or a documented self-service export of content and model together. 9 11 13
The Localisation Lead
The schema editor with models, fields, components, enumerations and reference fields is the reusable structure I want, webhook payloads carry localized fields per entry under a localizations key, and five named translation connectors (Lokalise, Crowdin, DeepL, EasyTranslate, Smartling) cover the German-English-French cycle. What keeps it short of the top is that we found no public information on translation status per locale, fallback languages, or a documented full export of content, model and locales with re-import in both directions. 9 10 11 12
The Web Engineer
The schema editor gives models, fields, components and enumerations, with reference fields linking entries so the model is more than forms, and localised fields ride along in webhook payloads under a localizations key. Named translation connectors — DeepL, Lokalise, Crowdin, Smartling, EasyTranslate — are documented, and AI Assist translates into target locales. We found no public information on a documented, self-service export of model, content, locales and assets, or on migration tooling or a CLI in either direction, which for a headless claim is the portability test. 9 10 11 12
The Data Protection Officer
The schema editor with models, fields, components and enumerations, reference fields between entries, and localized fields carried whole in webhook payloads give the repository genuine structure and reuse, with translation connectors named for Lokalise, Crowdin, DeepL, EasyTranslate and Smartling. We found no public information on fallback languages, translation status per locale, or a documented self-service export of content, model and assets together, which is what a migration buyer needs. 9 10 11
The Skeptic
The content model is genuinely structural — models, components, enumerations, reference fields between models, environments for testing schema changes — with localization connectors named and localized fields carried in webhook payloads. But we found no public information on field-level locale fallback, a media library with alt text, or any documented export of entries, model and locales, and portability is the part I read for: nothing captured shows the repository leaving in a format that includes locales and history. 9 10 11
Publishing, distribution & integrations
Show reasoningHide reasoning
How this is scored
Getting approved content to where it is read: publishing to websites and channels, a documented API and webhooks, and named connections to the CMS, CRM, analytics, SEO and translation tools a marketing stack already runs.
0 — Content leaves only by copy and paste; no API, no webhooks and no named integrations.
3 — Publishing to one CMS through a plugin, a handful of named integrations, and an API that is mentioned but not publicly documented.
5 — Publishing to websites and several channels from one place, a publicly documented REST or GraphQL API, webhooks on publish and change, and integrations named for at least the common CMS, CRM and analytics tools.
8 — A versioned API with published rate limits and SDKs, preview environments, a marketplace or integration directory with maintained connectors, performance data per piece pulled back from analytics, and a public status page.
10 — The platform is a dependable source for every channel: API deprecation policy published, CDN and delivery SLA stated, scheduled and event-driven publishing documented end to end, and content performance attributed per channel without a separate tool.
The Head of Content
A GraphQL-native API with published rate limits per plan and genuinely well-documented webhooks — signatures, retries, logs, pause and resume — make delivery dependable, and the marketplace names connectors for commerce, DAM, hosting, localization and PIM. We found no public information on named CRM or analytics connectors, SDKs, or performance data pulled back per piece. 1 2 10 11
The SME Marketer
Webhooks are documented in impressive depth — triggers per stage and action, retries, HMAC signatures, logs — rate limits are published per plan, and the marketplace names connectors across commerce, digital assets, hosting, localisation and product information. We found no public information on SDKs, per-piece analytics data pulled back, a delivery SLA or an API deprecation policy, which leaves the strongest delivery assurances unevidenced. 1 2 10 11
The Localisation Lead
A GraphQL-native documented API with published rate limits per plan, webhooks with HMAC signatures, retries and logs, plus a marketplace of maintained connectors across commerce, digital assets, hosting and localization make delivery solid, with Live Preview and an API status link listed. We found no public information on SDKs, analytics connectors pulling performance data back per piece, or a delivery SLA and API deprecation policy. 1 2 10 11
The Web Engineer
This is the strong suit: webhooks fire on publish, unpublish, create, update and delete, are environment-scoped, HMAC-SHA256 signed with a gcms-signature header, carry a published retry policy (4 retries on top of the initial request) and logs, and rate limits of 5, 25 and up to 500 requests per second for uncached requests are public per plan. A Marketplace with maintained connectors spans commerce, DAM, hosting, PIM and localisation, Live Preview is a compare-table row, and an API Status link sits in the navigation. We found no public information on a versioned API or deprecation policy, SDKs, or analytics connectors pulling performance data back per piece. 1 2 10 11
The Data Protection Officer
Webhooks are documented end to end — publish and unpublish triggers, HMAC-SHA256 signatures, a retry policy, logs with 7-day retention — rate limits per plan are published, an API status page is linked, and a marketplace carries named connectors for commerce, assets, hosting, localization and PIM. We found no public information on SDKs beyond a webhook signature utility, per-piece performance data returned from analytics, an API deprecation policy, or a delivery SLA. 1 2 10 11
The Skeptic
Webhooks are documented to a level few vendors reach — triggers by stage and action, HMAC-SHA256 signatures, a stated retry policy, logs — alongside a marketplace of named commerce, DAM, hosting and localization connectors, a GraphQL playground, and published rate limits per plan. Commerce and localization are well covered, but we found no public information on named CRM or analytics connectors in the marketplace, performance data pulled back per piece, or a delivery SLA. 1 2 10 11
AI assistance & content rights
Show reasoningHide reasoning
How this is scored
What the platform's AI does with the customer's content and what the customer may do with the AI's: the models and where they run, whether content trains them, opt-out, labelling of generated text, and who holds the rights to output.
0 — AI writing features with no statement on which model is used, where content is sent, or whether it is used for training.
3 — An AI page that names a provider but leaves open whether customer content trains models, where prompts are processed, or whether the features can be switched off.
5 — The model providers named as subprocessors, customer content excluded from training stated in the terms, AI features switchable per workspace by an admin, and ownership of generated output assigned to the customer.
8 — EU processing for AI features stated or selectable, retention of prompts and outputs given in days, generated or AI-edited content markable for EU AI Act transparency, and brand voice or terminology rules applied without uploading content to train a shared model.
10 — AI is governed and documented: per-feature list of model, provider and processing location, admin controls per role and content type, an audit trail of AI-generated changes, and the output rights and any indemnity written into the contract terms.
The Head of Content
AI Assist is documented down to supported field types, review-before-apply and token budgets on the billing page, and admin-managed guidelines apply brand, legal, glossary and audience context to every AI task. But we found no public information on which model is used, where prompts are processed, whether customer content is used for training, an opt-out or kill switch, or explicit rights to generated output beyond the general terms on customer content. 2 5 12 13
The SME Marketer
I write half our content with an assistant, and the AI Assist pages describe tasks, a review panel and brand guidelines in detail, yet we found no public information on which model or provider is used, where prompts are processed, or whether my drafts are used for training; the terms only confirm I keep rights I already hold in my content. Access appears to follow editor permissions with no additional role configuration, and we found no public information on labelling generated text, prompt retention, or switching the features off. 5 12 13
The Localisation Lead
AI Assist's generate, improve and translate tasks, review-before-apply, editable results and admin-managed guidelines are documented, with token budgets per plan and usage visible on the billing page. But the captured pages give no statement on which models run AI Assist, where prompts are processed, whether customer content trains them, or how features are switched off per workspace, and the terms address rights in customer content without addressing rights in AI output. 2 5 12 13
The Web Engineer
The AI pages are genuinely good on workflow control — suggestions shown in a review panel before anything is written, applied changes editable like any manual edit, admin-only guideline management with reusable brand-voice and glossary context, token usage visible on the billing page. But we found no public information on which model or provider is used, where prompts are processed, whether customer content trains models, retention of prompts and outputs, or EU AI Act labelling of generated text. The terms state you retain the rights you already hold in your content; the questions a buyer must answer before sending content to an AI are left open. 5 12 13
The Data Protection Officer
AI Assist is described functionally — generate, improve and translate, with a review panel before anything is written, token allowances of 100.000, 300.000 and 2.000.000+ by plan and usage visible on the Billing page — and admin-managed guidelines cover brand voice, legal compliance and glossary. We found no public information on which model or provider processes prompts, where that processing happens, prompt and output retention in days, or an exclusion of customer content from training written into the terms, and the terms assign rights only to content the customer already holds. 2 5 12 13
The Skeptic
AI Assist is documented as a product — generate, improve and translate tasks, suggestions reviewed before anything is written to an entry, admin-managed guidelines, token usage visible on the billing page. On the rights side, we found no public information on which model or provider is used, where prompts are processed, whether customer content trains the models, retention of prompts and outputs, or how an admin switches the features off; the terms return only rights the customer already held rather than assigning rights to generated output. 5 12 13
European sovereignty
panel opinion
panel disagrees
Show reasoningHide reasoning
How this is scored
Who the contracting entity is, where content, assets and AI prompts are stored and processed, and which subprocessors touch them. Independently sourced by the sovereignty pipeline.
0 — Non-EU contracting entity, no statement on where content is hosted, and subprocessors unnamed.
3 — Content hosted in the EU on request or on a higher plan only, or the subprocessor list is absent, or AI and analytics features send content outside the EU without an explained safeguard.
5 — EU hosting stated for content and assets, a DPA available, but the contracting entity or significant parts of the chain (CDN, AI, support) are non-EU without a stated safeguard.
8 — EU contracting entity, content, assets and backups stored in the EU on named infrastructure, subprocessor list published with locations and purposes, and AI processing location stated.
10 — Sovereign end to end and evidenced: European ownership, EU entity, every hosting, CDN, AI and support subprocessor European or optional, self-hosting or a European cloud offered, and certifications published.
The Head of Content
The imprint places the contracting entity in Berlin, which is the right start for a European buyer. We found no public information on where content, assets and AI prompts are stored and processed, and no published subprocessor list with locations and purposes — the captured privacy statements cover website visitor data rather than customer content. 1 3 4
The SME Marketer
The vendor presents itself as a Berlin company, yet nothing captured states where content, assets or AI prompts are stored or processed, and we found no public information on a subprocessor list; the independently computed sovereignty attributes record entity, ownership, residency and subprocessor exposure as unconfirmed on the vendor's own pages. For a German Mittelstand buyer, that is not enough to sign. 3 5 12
The Localisation Lead
The privacy policy reads GDPR-German — double opt-in, server logs, deletion when the storage purpose ends — but it addresses website visitors, and we found no public information on where content, assets and AI prompts are hosted, on a DPA, or on a published subprocessor list with locations and purposes. For a Berlin-based vendor this is the page a DACH buyer needs first, and the captured pages do not carry it. 3 5
The Web Engineer
The provenance record names Hygraph GmbH in Berlin as the contracting entity, and the privacy policy reads GDPR-style with statutory retention and deletion duties. But we found no public information on the captured pages confirming where content, assets and backups are hosted, a subprocessor list with locations and purposes, or where AI prompts are processed — an EU entity with the rest of the chain unknown. 3 4
The Data Protection Officer
The privacy policy runs on German-style procedures such as double opt-in and statutory blocking or deletion, but we found no public information confirming where content, assets and backups are hosted, no data processing agreement, no published subprocessor list with locations and purposes, and no statement on where AI prompts are processed. For content this sensitive, the contracting entity and hosting region should be named on the vendor's own pages, and these captures do not do it. 3 4 12
The Skeptic
The sovereignty record as independently sourced is empty across jurisdiction, ownership, residency and subprocessors, so I score it at the floor the anchors define for that state. We found no public information on where content, assets and backups are hosted, which subprocessors touch them, or where AI prompts are processed; the privacy policy that was captured concerns website cookies and server logs rather than content residency. 3 12
Pricing transparency
panel disagrees
Show reasoningHide reasoning
How this is scored
Whether a buyer can compute the annual cost for their team, locales and traffic from public pages alone — seats and contributor roles, locales, content records, API calls and bandwidth, AI credits, and the plan where approval workflows or SSO begin.
0 — No public prices at all; every plan is a sales conversation.
3 — A starting price per month, but what it includes is vague and the limits that drive the real cost (locales, records, API calls, AI credits) are unstated.
5 — Plans priced per seat with their main limits listed, but at least one commonly incurred cost (extra locales, overage on records or API calls, AI credits, guest contributors) is missing or "on request".
8 — Every plan priced with seats, locales, records, API and bandwidth limits and overage rates stated, AI usage priced or included explicitly, and the plan that unlocks approval workflows, SSO and audit logs named.
10 — Complete cost computability: annual cost derivable for a given team, locale count and traffic, every add-on and overage priced publicly, contract term and renewal conditions stated, and the cost of exporting or leaving the platform zero or published.
The Head of Content
The public plan comparison states concrete limits — AI tokens, API tokens, rate limits for uncached requests, webhooks and content stages — and names the read replica as an enterprise add-on. We found no public information on the monetary price of any plan, seat or contributor costs, locale or record limits, overage rates, or the plan where approval workflows and SSO begin, so I cannot compute an annual cost from what is captured. 2
The SME Marketer
The plan comparison is unusually concrete about limits — rate limits, webhooks, API tokens, AI tokens and content stages per plan — and states that third-party tools carry no extra Hygraph fee but bill separately. We found no public information on plan prices, seat or guest costs, extra locales or AI overage rates, so I cannot compute what a second language and the agency as guests would cost. 2 12
The Localisation Lead
The compare table publishes unusually concrete limits — AI tokens 100.000, 300.000 and 2.000.000+, API tokens, rate limits of 5, 25 and up to 500 uncached requests per second, webhooks and content stages — plus monthly versus yearly billing and a Read Replica add-on for enterprise. The captured pricing page gives no price figures, and we found no public information on locale or record limits, AI overage rates, or the plan where approval workflows and SSO begin, so an annual cost is not computable from the capture. 2
The Web Engineer
The compare table publishes real limits per plan — rate limits for uncached requests, API tokens, webhooks, content stages, AI tokens at 100.000, 300.000 and 2.000.000+, and support channels — and AI token usage is visible on the billing page. But we found no public information on seat prices or any monetary figures, record or locale limits, overage rates, or the plan that unlocks approval workflows, SSO and audit logs, so an annual cost is not computable from the captured pages. 2 12
The Data Protection Officer
The plan comparison publishes concrete limits — API tokens, webhooks, uncached request rate limits and content stages per plan, monthly and yearly billing options, and a note that third-party tools bill separately with no fee from Hygraph. We found no public information on seat or contributor pricing, locale or content record limits, overage rates, or a figure for the Read Replica beyond the word "Add-on", so an annual cost for a given team cannot be computed from these pages. 2
The Skeptic
The plan comparison publishes real limits — API tokens, requests per second for uncached requests, webhooks, content stages, AI tokens per plan — which is more limit disclosure than most in this category. But not a single price appears in the captured pages, and we found no public information on seat pricing, locale limits, overage rates for records, API calls or AI tokens, or the plan where approval workflows and SSO begin, so an annual cost is not computable from what is public. 2
European sovereignty — proven facts
0 of 4 dimensions provenBuilt only from facts shown on the vendor's own pages. A dimension we could not prove is left open, not scored as zero.
| Legal entity | Not determined | — | uncited Report an error |
|---|---|---|---|
| Ownership | Not determined | — | uncited Report an error |
| Data residency | Not determined ⚠ unverified | — | uncited Report an error |
| Subprocessors | Not determined ⚠ unverified | — | uncited Report an error |
Questions about Hygraph
Is Hygraph European?
Yes. whats-best.ai classifies Hygraph as EU-Made.
Where does Hygraph store data?
Not stated on the vendor’s pages as of 13 Oct 2026.
Does Hygraph have a free plan?
Yes. Hygraph has a free plan, per the vendor’s page captured 13 Oct 2026.
What are the best alternatives to Hygraph?
As of 15 Oct 2026, the panel rates CoreMedia (3/5), Storyblok (2.5/5) and Contently (2.5/5) highest among the other 14 Content Marketing products evaluated. Ratings that look equal are rounded to half stars; the order uses the unrounded rating, then the name. The highest-rated European alternative is CoreMedia (EU origin, foreign-owned). Ratings are the AI panel’s opinion, formed from public evidence: how scores are made. The full list, with European products first: Hygraph alternatives.
Where this could be wrong
- Evidence ages. The oldest capture behind this page is from 13 Oct 2026. Vendors change pricing and policies without notice; every fact reflects its source as of the capture date shown in the registry.
- Weak sourcing — Data residency, Subprocessors. Not confirmed on the vendor’s own pages as captured.
- AI can misread a source. Extraction and judgement are automated; a citation guarantees traceability, not infallibility. If something here is wrong, say so — no account needed, every report is decided within 5 business days, and accepted corrections are published.
What we left out
A claim that does not survive our checks costs us the claim, not the page. This is what was taken off this one.
- 27 product facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 6 subprocessors facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 3 compliance facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 3 legal facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 2 data facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 2 pricing facts could not be confirmed on the vendor’s page as captured and were left out of this page and of the panel’s material. Know more? Tell us
- 2 sovereignty dimensions could not be confirmed on the vendor’s own pages and are shown as unknown. Know more? Tell us
- 6 of the readings below were written against an earlier fact sheet — a fact has been corrected, added or pulled since. Until the panel next runs on this product you are reading the older judgement. Know more? Tell us
Sources (13)
The pages every claim on this page was read from — each one checked, dated, and kept verifiable.
- 1 Vendor homepage hygraph.com Checked 13 Oct 2026 Details →
- 2 Pricing page hygraph.com Checked 13 Oct 2026 Details →
- 3 Privacy policy hygraph.com Checked 13 Oct 2026 Details →
- 4 Legal notice hygraph.com Checked 13 Oct 2026 Details →
- 5 Terms of service hygraph.com Checked 13 Oct 2026 Details →
- 6 Creation, review & approval — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 7 Creation, review & approval — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 8 Structured content, localisation & export — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 9 Structured content, localisation & export — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 10 Publishing, distribution & integrations — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 11 Publishing, distribution & integrations — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 12 AI assistance & content rights — found from sitemap hygraph.com Checked 13 Oct 2026 Details →
- 13 AI assistance & content rights — found from sitemap hygraph.com Checked 13 Oct 2026 Details →