Storyblok

Connect AI agents to the Storyblok headless CMS. Agents read published or draft stories, datasource entries, links, and tags through the Content Delivery API, and create, update, publish, and delete stories through the Management API.

12 actions available

A batch of content rows or a form submission arrives

Agent maps each row to a story with the right component

Agent confirms the slug and required content fields are set

Agent stages the story with Create Story without publishing

The editorial channel gets a note that drafts are ready

An editor approves before the agent publishes the story live

What This Integration Enables

Through the Content Delivery API, an agent reads stories with filtering by slug prefix, draft or published version, sorting, and field filtering, gets a single story by slug or id, and reads datasource entries, tags, and the space's link tree. The link tree is a lightweight map of every story with its slug, parent, and folder flag, which is ideal for building navigation menus and sitemaps. Reading the draft version requires a preview token. Through the Management API, the agent creates, updates, publishes, and deletes stories, and lists stories with editorial metadata such as unpublished changes. It can also read space details and list assets. Story content must include a component key that matches a content type defined in the space. This lets an agent draft content from spreadsheets, forms, or generated drafts and stage it for review, while the publish step that makes content live stays with a person.

Without FlowRunner

Manual story creation Someone keys each content row into Storyblok by hand

Publishing has no gate Stories go live whenever an editor gets to them

Navigation built by hand Menus and sitemaps are maintained manually as content changes

With FlowRunner

Stories staged by agent The agent creates each row as a draft story with its component set

Publishing is reviewed An editor signs off before the agent publishes the story

Navigation from the link tree The agent builds menus from the space's link tree automatically

Use Case Scenarios

Load a Content Batch

The agent reads a batch of rows from [Google Sheets](/integrations/google-sheets), then calls Create Story for each row to load it into the space as a draft with the right component. The team gets a queue of ready stories instead of a spreadsheet to copy from.

Draft From a Form Submission

When a form is submitted on the marketing site, the agent calls Create Story to draft a matching content entry, then holds it for review before Publish Story makes it live. The submission becomes a structured draft the moment it comes in, without going public unreviewed.

Notify on New Drafts

The agent uses Get Stories to fetch recently updated draft content, then posts to the editorial channel with [Slack](/integrations/slack) Send Message To Channel so the team knows new drafts are ready to review. Nobody has to keep checking the CMS to find what changed.

Human-in-Loop Highlight

Reading content, drafting stories, and building navigation are safe to automate, so the agent runs them on its own. Publishing is the consequential step, because Publish Story makes the current draft content live for every site and app reading the space. So when a flow reaches the publish step, the agent stops. It routes the drafted story to an editor as a [human-in-the-loop](/concepts/human-in-the-loop/) step, showing the story name, the full slug, the component, and a preview of the content, and asks: "Publish this story to the live space, or hold for edits?" On approval the agent calls Publish Story. On hold, the story stays a draft and nothing goes live. The AI does the drafting; the decision to publish gets a person.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

12 actions

Content Delivery

Get Stories Retrieves a paginated list of stories with filtering by slug prefix, draft or published version, sorting, and advanced field filtering. The draft version requires a preview token.
Get Story Retrieves a single story by its full slug or numeric id, including its resolved content. The draft version requires a preview token.
Get Datasource Entries Retrieves key and value entries from a datasource, useful for shared configuration values, translation strings, and select options.
Get Links Retrieves the link tree of a space, a lightweight map of all stories with slugs, parent relationships, and folder flags, ideal for building navigation menus.
Get Tags Retrieves all tags used across stories in a space, each with a count of how many stories use it.

Content Management

Create Story Creates a new story with a display name, a full slug, and a content object whose component matches a content type in the space. Can publish on creation.
Update Story Updates an existing story, changing only the fields you provide. Can publish the story after updating.
Delete Story Permanently deletes a story from the space. This cannot be undone.
Publish Story Publishes a story, making its current draft content live.
List Stories Lists stories including drafts and unpublished content, with pagination and slug-prefix filtering. Use this when you need editorial metadata such as unpublished changes.
Get Space Retrieves details about the space, including its name, plan, domain, default language, and environment settings.
List Assets Lists assets in the space, including images, documents, and other uploaded files, with pagination and folder filtering.

Start building with Storyblok

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us