For Strapi and headless sites
Blog AI for Strapi
For teams whose marketing site runs on Strapi, the open-source headless CMS. Blog AI finds what to write from your Search Console data, drafts posts in your voice with checked sources, and publishes them into your articles collection with the SEO title, description, canonical, featured image, category and tags filled in.
You connect with an API token. No plugins to install, and your front end stays yours.
What Blog AI writes into Strapi
Blog AI publishes to a collection with the API name articles. Only three fields are required. The rest are filled when your schema has them, and skipped with a warning when it doesn’t.
| Field | What goes in it |
|---|---|
| title, slug, body | Required. The body is Markdown. The slug is set on first publish and never changed afterwards. |
| excerpt | A short summary, up to 300 characters. |
| seoTitle, seoDescription, meta_description | The title and description search engines show, trimmed to 60 and 160 characters. |
| canonicalUrl | The address search engines should treat as the original, when the post has one. |
| thumbnail, ogImage | The featured image, uploaded to your media library with its alt text, used for the post and for social previews. |
| category, tags | Matched by name and created when missing. Only the first category is sent, because the field links each article to one category. |
Connect Strapi in five steps
Step 1: Create an API token in Strapi
In your Strapi admin, go to Settings, then API Tokens, and create a token. A custom token is best: give it find, findOne, create, update and delete on articles, categories and tags, plus upload for images. A full-access token also works. Copy the token when Strapi shows it.
Step 2: Add the site in Blog AI
Sign in to Blog AI with Google, open Sites and choose Strapi. Enter a name, your public website address (where readers see the blog), the address of your Strapi server without /api at the end, and the API token.
Step 3: Set the post address pattern
Tell Blog AI where each post appears on your website, with {slug} where the post's slug goes, for example https://example.com/blog/{slug}. If you leave it empty, Blog AI uses your website address followed by /blog/{slug}.
Step 4: Test the connection
Blog AI reads your articles collection with the token. If the token is refused, the collection is missing or the address redirects, the test says which, so you can fix it before anything is written.
Step 5: Import your posts (and, optionally, your pages)
Blog AI imports your published articles to learn your voice and link to them. If your pages live in their own collection, add its name under Fix connection to import them read-only for the content audit.
Why headless sites fit Blog AI well
A headless CMS keeps content separate from the website that shows it. That’s good for developers and awkward for marketers: every post needs its SEO fields, image and relations filled in by hand in the admin panel, and nobody checks whether it worked.
Blog AI fills those fields for you and then measures the result. Because each post has a known public address (from your post address pattern), its clicks, impressions and visits show on the post page with a marker for the day it was published or refreshed. The same goes for posts you imported: once they’re matched, the content audit gives each one a recommended action.
Our own blog reads from Strapi, and we’re moving its publishing onto this same connector, as a normal Blog AI customer. The connector was built against the article schema we use ourselves, so if you’re starting a new Strapi blog, copying that shape (the fields in the table above) gives you every field Blog AI can fill.
After you connect
Your voice, from your own articles
Blog AI imports your published articles, whether their body is Markdown or Strapi’s blocks format, with their relations and cover image. With 3 or more, it learns your voice. Posts Blog AI published itself are never overwritten by a later import.
Pages for the audit (optional)
Strapi has no standard pages type, so Blog AI doesn’t guess. If your pages live in a collection such as pages, add its API name under Fix connection on the Sites page. Blog AI checks that it can read it, then imports the pages read-only: their text (from a body or content field, or the text inside block and section zones) and SEO fields. Pages are never written back to Strapi.
Updating live posts
Edit a published post in Blog AI and click Update live post to send the change to Strapi. When you accept a refresh, Strapi is updated first and Blog AI saves its copy after. If someone changed the article in Strapi since Blog AI last sent it, Blog AI stops and asks you to re-import rather than overwrite their edit.
Scheduling
Strapi’s API has no “publish later”, so scheduled posts wait in Blog AI and are created in Strapi at the time you pick, within a few minutes. They show on the Blog AI calendar until then, not in your Strapi admin.
What it costs
The connection is free: importing, publishing, scheduling and updating cost nothing. Credits pay for the AI work.
| Action | Credits |
|---|---|
| A batch of post ideas | 1 credit |
| A researched, fact-checked blog post | 3 credits |
| A generated image | 1 credit |
| A post refresh (rewrite of the parts that need it) | 2 credits |
| A full rewrite of a post in your site's voice | 3 credits |
| Merging overlapping posts into one | 3 credits |
| Better title & description in Google (when your search data shows it's needed) | Free |
| Publishing, scheduling, editing, uploads | Free |
What it doesn’t do
- It doesn’t change your content types, roles, plugins or front end.
- It doesn’t rebuild or redeploy your website. If your site caches Strapi content, it shows new posts on its usual schedule.
- It publishes to the articles collection only; it never writes to your pages.
- It doesn’t publish anything until you publish or schedule it.
- It can’t read a Strapi server that isn’t reachable over https from the internet.
Questions about Strapi
Which versions of Strapi work?
The connector is built for Strapi 5 and its REST API with an API token. If you run an older version, contact us before you connect so we can tell you what to expect.
Do I have to change my content types?
Not usually. Blog AI needs an articles collection with a title, a slug and a Markdown body. Every other field is optional: if your schema doesn't have one, Strapi says so, Blog AI leaves that field out, shows you a warning, and publishes the rest.
My posts collection isn't called articles. Can I still connect?
Blog AI publishes to a collection whose API name is articles. If yours is named differently, contact us before connecting and we'll tell you your options.
How does scheduling work without native scheduling?
Strapi has no built-in way to publish at a future time through its API, so Blog AI holds a scheduled post and creates it in Strapi when the time comes. Scheduled posts go out within a few minutes of the time you pick. If the post fails, you get an email with the reason.
Will my website show the new post straight away?
That depends on your front end. Blog AI writes to Strapi; your site shows the post when it next fetches from Strapi. If your site is statically built or cached, you may need a rebuild or revalidation hook in Strapi, as you would for a post written by hand.
What does the Strapi connection cost?
Connecting, importing, publishing, scheduling and updating live posts cost nothing. Writing a new post costs 3 credits; the table on this page lists every paid action.
Start with the data you already have
Sign in with Google, no card needed. The free plan includes a content audit and one free post once Search Console has synced and you've set your voice. Plans from $49/month.