What is the Storyblok MCP server?
The Storyblok MCP server is a community Python server built on FastMCP that lets an MCP client manage a Storyblok space. Its README lists tools for stories, components, assets, datasources, tags, releases, workflows and webhooks, including bulk operations, under an MIT licence.
It is a modular Model Context Protocol server written in Python with FastMCP and httpx. Each Storyblok resource has its own tool module, so an agent can manage a space through the Management API without you clicking through the app.
The README describes four verbs across stories, components, assets, datasources, tags, releases and workflows: create, read, update and delete. A meta tool lets the agent list every available tool at runtime.
What an agent can do with it
- Fetch, create, update, publish and delete stories, including bulk versions of each.
- Manage components, folders, presets and field plugins, and find which stories use a component.
- Handle assets, asset folders and upload steps.
- Work with releases, scheduled stories, workflows, stages and approvals.
- Run an AI translation of a story and compare story versions.
Who it is for
Best for
- Developers and editors with a large Storyblok space and repetitive content chores.
- Teams that want an assistant to audit component usage before changing a schema.
- Marketers who would rather ask for a list of stories than filter in the app.
Not for
- Anyone wanting minimal permissions. The README describes deleting stories in bulk and managing collaborators.
- Users who need a hosted server. It runs locally from a clone.
Use it with UiChemy
Storyblok is a headless CMS, so the page design lives elsewhere. For a WordPress front end, UiChemy's Figma to WordPress converts designs into editable pages. Click Get Started on uichemy.com.
Connect Storyblok MCP Server: Manage Spaces and Stories With AI to your AI client
git clone <your-repo-url> && cd storyblok-mcp-server && pip install -r requirements.txtClone and install
Clone the repository and run pip install -r requirements.txt. The README notes uv must also be installed.
Fill in credentials
Copy .env.example to .env and set the space ID, a management token and a default public token.
Add the client config
Use the uv based block above, pointing the last argument at server.py on your machine.
Check the tool count
Restart the client and confirm the tools appear. You can also run mcp run server.py with the MCP Inspector.
Try a safe prompt
The README suggests asking for all stories or the details of one story before attempting changes.
Client configuration
{
"mcpServers": {
"storyblok": {
"command": "uv",
"args": ["run", "--with", "mcp", "mcp", "run", "/path/to/storyblok-mcp-server/server.py"],
"env": {
"STORYBLOK_SPACE_ID": "your_space_id",
"STORYBLOK_MANAGEMENT_TOKEN": "your_management_token",
"STORYBLOK_DEFAULT_PUBLIC_TOKEN": "your_public_token"
}
}
}
}Endpoints and ports
your_space_idThe space to manage.your_management_tokenManagement API token.your_public_tokenPublic access token.What you can do with it
- Broad API coverageThe README tables around thirty resource areas, from access tokens and approvals to workflows and webhooks.
- Modular tool designEach resource lives in its own module, which keeps the code easy to extend.
- Bulk operationsPublish, update, create or delete several stories at once.
- Component usage lookupFind which stories depend on a component before you edit it.
- Version historyList versions of stories and components, compare two versions and restore one.
- Tool discoveryA meta tool reports every available tool and its description at runtime.
Tools it exposes
fetch_stories / get_story
List stories with filtering or read one by ID.
create_story / update_story / delete_story
Write and remove single stories.
publish_story / unpublish_story
Change publish state.
bulk_publish_stories / bulk_update_stories / bulk_delete_stories
Act on many stories in one call.
get_story_versions / restore_story / compare_story_versions
Review and roll back history.
ai_translate_story
AI powered translation for a story.
fetch_components / get_component_usage
List components and find the stories that use one.
fetch_assets / init_asset_upload / complete_asset_upload
Browse and upload assets.
retrieve_multiple_activities
List activity log entries.
Frequently asked questions
What is the Storyblok MCP server?
It is a community Model Context Protocol server, written in Python, that exposes the Storyblok Management API as tools. An AI client can use it to manage stories, components, assets, datasources, tags, releases and workflows in a space. The repository is MIT licensed.
How do I install the Storyblok MCP server?
Clone the repository, run pip install -r requirements.txt and make sure uv is installed. Copy .env.example to .env with your space ID and tokens, then add a uv based entry in your MCP client config that points to server.py.
Which tokens does it need?
The README lists a space ID, a management token and a default public token as environment variables. The management token can change and delete content, so create it for a test space first and store it only in your local environment file.
Can it publish and delete stories in bulk?
Yes. The README lists bulk publish, bulk update, bulk create and bulk delete tools for stories. Because a single request can change many stories, preview the list with a fetch first and keep a content backup.
Does it include AI translation?
The README lists an ai_translate_story tool described as AI powered translation for a story. It gives no detail on the model used or any cost, so check how it behaves on a draft story before relying on the output.
Is it the official Storyblok MCP server?
No. It is a community project by Kiran1689, and the README does not claim any affiliation with Storyblok. Check Storyblok's own documentation if you need a vendor supported connector, and compare features first.
How active is the project?
The repository was last pushed in June 2025 and shows 41 stars. That is a long gap, so read the open issues, confirm the tools still match the current Management API, and test every call you need on a copy of your space.
What else could my web builder connect to?
For WordPress, look at Respira for WordPress across many page builders, or MCP Abilities Elementor. For design files, the Figma MCP server is the usual link. Pick by whether your problem is content, page building or design handoff.


