Paste your endpoint and one token — articles arrive as published entries on your own Hygraph model.
Hygraph is GraphQL-native and built for programmatic content. You paste your Content API endpoint and a Permanent Auth Token, DraftSEO reads your schema, and you confirm which fields hold the title, the body, the slug and the cover image. Each article is then created on your own model and published for you — Hygraph keeps drafts and published entries separate, so DraftSEO always does both steps. Every image is ingested as a Hygraph asset and published too, because an unpublished asset will not deliver. Hygraph is headless, so your own website renders the article; you tell DraftSEO the address pattern your posts live at.
| Connection | Endpoint + Permanent Auth Token — Project Settings → API Access |
|---|---|
| Body formats | RichText · plain text — matched to your body field |
| Image hosting | Hygraph assets on graphassets.com — ingested, never hot-linked |
| Publishing | Created and published in one run — Hygraph has no create-and-publish, so DraftSEO does both |
| Your own fields | Left untouched on republish — only the mapped fields are written |
| Publish modes | Auto · draft · scheduled |
Who does what
Nothing — once the endpoint and token are in and the fields are mapped, entries appear on your model already published, and your website shows them. Want to review first? Send them as drafts and press Publish in Hygraph.
Works with whatever your model and its fields are called.
Setup
In Hygraph open Project Settings → API Access → Permanent Auth Tokens → Add Token. Grant Create, Update, Publish and Read on Published — and Read on the Draft stage — for your article model and for Assets.
~2 minutes
On the same API Access screen, copy the endpoint exactly as shown. DraftSEO uses the address your project gives you rather than guessing your region.
~30 seconds
Connect a site → Hygraph → paste the endpoint and the token. DraftSEO reads your schema and lists the models an article can go into.
~1 minute
Confirm which fields hold the title, body, slug and cover image, then tell DraftSEO where posts appear, e.g. https://www.yoursite.com/blog/{slug}.
~2 minutes
FAQ
It is the single most common Hygraph connection failure. Creating an entry writes to the Draft stage, but the same request reads the result back from Published by default — so a token that cannot read Draft fails with "Mutation failed due to permission errors" even though everything else is correct. Grant Read on Draft for your model and for Assets and it works. DraftSEO checks this when you connect and tells you if it is missing.
Your choice. Hygraph has no single create-and-publish step, so DraftSEO creates the entry and then publishes it for you. Choose draft instead and the entry is created and left unpublished for you to review — if the article was already live, its published version stays as it was rather than being pulled down.
Yes. DraftSEO converts each article into Hygraph RichText with headings, lists, links and images. Plain text and String fields are supported too — DraftSEO matches whichever your body field actually is. Tables become one line per row, because a malformed RichText table breaks the renderer on your own site.
Each one is ingested by Hygraph and re-hosted on graphassets.com, then published — an unpublished asset would not appear on your site. DraftSEO never leaves an image pointing back at its own storage. If an image cannot be ingested it is left out rather than published broken, and DraftSEO retries it in the background.
No — it works on the free plan. Bear in mind that assets count towards the free plan entry limit, so an illustrated article uses roughly two entries.
No card required. Images ingested as Hygraph assets.