Storyblok MCP Server
# Storyblok MCP Server
[](LICENSE)
[](https://nodejs.org/)
[](https://modelcontextprotocol.io/)
[](https://www.typescriptlang.org/)
<a href="https://glama.ai/mcp/servers/martinkogut/storyblok-mcp-server">
<img width="380" height="200" src="https://glama.ai/mcp/servers/martinkogut/storyblok-mcp-server/badge" />
</a>
A TypeScript [Model Context Protocol](https://modelcontextprotocol.io/) server for the Storyblok Management API. It lets AI assistants and agents manage Storyblok content, workflows, and configuration safely and programmatically.
## Why this project
Storyblok is powerful, but repetitive content operations and release workflows can be time-consuming. This server bridges MCP-compatible assistants with Storyblok so teams can:
- Automate content ops (bulk updates, tagging, assets)
- Orchestrate releases and workflows with less manual effort
- Keep component libraries and schemas consistent
- Give AI agents structured, typed access to Storyblok
## Highlights
- **160 tools across 30 modules** mapped to Storyblok Management API capabilities
- **Typed schemas** with Zod for safer tool calls
- **MCP-first design** for MCP-compatible clients
- **Simple config** using environment variables
## What it does in 15 seconds
**Prompt**
```text
Find all stories updated in the last 7 days and tag them as "reviewed".
```
**Response (example)**
```text
Found 12 stories updated since 2025-12-28.
Applied tag "reviewed" to 12 stories.
```
## Quick Start
```bash
# Clone the repository
git clone https://github.com/hypescale/storyblok-mcp-server.git
cd storyblok-mcp-server
# Install dependencies
npm install
# Build
npm run build
```
Create a `.env` file from the example and fill in your credentials:
```bash
cp .env.example .env
```
Then start the server:
```bash
npm run start
```
## Requirements
- Node.js >= 20
- A Storyblok space with Management + Public/Preview tokens
## Configuration
The server requires three environment variables (see `.env.example`):
```env
STORYBLOK_SPACE_ID=your_space_id
STORYBLOK_MANAGEMENT_TOKEN=your_management_token
STORYBLOK_DEFAULT_PUBLIC_TOKEN=your_public_token
```
What they are used for:
| Variable | Description |
|----------|-------------|
| `STORYBLOK_SPACE_ID` | Your numeric Storyblok space ID |
| `STORYBLOK_MANAGEMENT_TOKEN` | Management API token with appropriate permissions |
| `STORYBLOK_DEFAULT_PUBLIC_TOKEN` | Public/Preview token for content delivery |
Tip: Use the smallest permission set possible for safety.
## Usage with MCP Clients
This server is designed to work with any MCP-compatible client. Here are ready-to-copy configs for popular tools.
### Claude Desktop
<details>
<summary>Show config</summary>
Add to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"storyblok": {
"command": "node",
"args": ["/path/to/storyblok-mcp-server/dist/index.js"],
"env": {
"STORYBLOK_SPACE_ID": "your_space_id",
"STORYBLOK_MANAGEMENT_TOKEN": "your_management_token",
"STORYBLOK_DEFAULT_PUBLIC_TOKEN": "your_public_token"
}
}
}
}
```
</details>
### Codex CLI / IDE
<details>
<summary>Show config</summary>
Codex supports MCP servers via CLI or by editing `~/.codex/config.toml`. The CLI and IDE extension share the same config.
Option A: Configure with the Codex CLI
```bash
codex mcp add storyblok \
--env STORYBLOK_SPACE_ID=your_space_id \
--env STORYBLOK_MANAGEMENT_TOKEN=your_management_token \
--env STORYBLOK_DEFAULT_PUBLIC_TOKEN=your_public_token \
-- node /path/to/storyblok-mcp-server/dist/index.js
```
Option B: Configure with `~/.codex/config.toml`
```toml
[mcp_servers.storyblok]
command = "node"
args = ["/path/to/storyblok-mcp-server/dist/index.js"]
[mcp_servers.storyblok.env]
STORYBLOK_SPACE_ID = "your_space_id"
STORYBLOK_MANAGEMENT_TOKEN = "your_management_token"
STORYBLOK_DEFAULT_PUBLIC_TOKEN = "your_public_token"
```
</details>
### Cursor
<details>
<summary>Show config</summary>
Create a project config at `.cursor/mcp.json` (or use the global config at `~/.cursor/mcp.json`):
```json
{
"mcpServers": {
"storyblok": {
"command": "node",
"args": ["/path/to/storyblok-mcp-server/dist/index.js"],
"env": {
"STORYBLOK_SPACE_ID": "your_space_id",
"STORYBLOK_MANAGEMENT_TOKEN": "your_management_token",
"STORYBLOK_DEFAULT_PUBLIC_TOKEN": "your_public_token"
}
}
}
}
```
</details>
### Windsurf
<details>
<summary>Show config</summary>
Create `~/.codeium/mcp_config.json`:
```json
{
"mcpServers": {
"storyblok": {
"command": "node",
"args": ["/path/to/storyblok-mcp-server/dist/index.js"],
"env": {
"STORYBLOK_SPACE_ID": "your_space_id",
"STORYBLOK_MANAGEMENT_TOKEN": "your_management_token",
"STORYBLOK_DEFAULT_PUBLIC_TOKEN": "your_public_token"
}
}
}
}
```
</details>
### VS Code
<details>
<summary>Show config</summary>
Create `.vscode/mcp.json` in your workspace:
```json
{
"servers": {
"storyblok": {
"type": "stdio",
"command": "node",
"args": ["/path/to/storyblok-mcp-server/dist/index.js"],
"env": {
"STORYBLOK_SPACE_ID": "your_space_id",
"STORYBLOK_MANAGEMENT_TOKEN": "your_management_token",
"STORYBLOK_DEFAULT_PUBLIC_TOKEN": "your_public_token"
}
}
}
}
```
</details>
### Continue
<details>
<summary>Show config</summary>
Create a YAML config at `.continue/mcpServers/storyblok.yaml`:
```yaml
name: Storyblok MCP
version: 1.0.0
schema: v1
mcpServers:
- name: storyblok
command: node
args:
- /path/to/storyblok-mcp-server/dist/index.js
env:
STORYBLOK_SPACE_ID: your_space_id
STORYBLOK_MANAGEMENT_TOKEN: your_management_token
STORYBLOK_DEFAULT_PUBLIC_TOKEN: your_public_token
```
</details>
### Cline
<details>
<summary>Show config</summary>
Open Cline settings and use "Configure MCP Servers" to open `cline_mcp_settings.json`, then add:
```json
{
"mcpServers": {
"storyblok": {
"command": "node",
"args": ["/path/to/storyblok-mcp-server/dist/index.js"],
"env": {
"STORYBLOK_SPACE_ID": "your_space_id",
"STORYBLOK_MANAGEMENT_TOKEN": "your_management_token",
"STORYBLOK_DEFAULT_PUBLIC_TOKEN": "your_public_token"
}
}
}
}
```
</details>
## Available Tools
The tool surface mirrors Storyblok's Management API. Total tool count and modules are defined in `src/tools/index.ts`.
| Module | Tools | Description |
|--------|-------|-------------|
| Stories | 18 | CRUD, bulk operations, publishing, versioning |
| Components | 9 | Schema management, versioning, usage tracking |
| Assets | 9 | Upload, organize, bulk operations |
| Workflows | 6 | Workflow management and stages |
| Releases | 5 | Release scheduling and deployment |
| Datasources | 5 | Key-value data management |
| Tags | 5 | Content tagging and organization |
| Webhooks | 5 | Event notifications |
| Collaborators | 4 | Team management |
| Space Roles | 5 | Permissions and access control |
| ... | ... | And 20+ more modules |
## Example Prompts
Use these prompts in your MCP client to see quick wins:
- "List all stories updated in the last 7 days and tag them as 'reviewed'."
- "Create a release for next Friday and add the homepage and pricing stories."
- "Find unused components and tell me which stories still reference them."
- "Upload these assets and organize them into a new 'Campaign' folder."
## Development
```bash
# Run in development mode with hot reload
npm run dev
# Type check
npm run typecheck
# Build for production
npm run build
```
## Project Structure
- `src/index.ts` - MCP server entry point
- `src/tools/*` - Tool modules mapped to Storyblok API areas
- `src/types/*` - Shared API and Storyblok types
- `src/utils/*` - HTTP helpers and response formatting
- `src/config.ts` - Environment configuration and validation
## Troubleshooting
- `ConfigError: STORYBLOK_* is missing` - ensure `.env` is present and filled out.
- `401 Unauthorized` - verify the Management and Public/Preview tokens.
- `404 Not Found` - double-check the `STORYBLOK_SPACE_ID`.
## Tech Stack
- TypeScript with strict type checking
- [@modelcontextprotocol/sdk](https://github.com/modelcontextprotocol/typescript-sdk) for MCP implementation
- Zod for runtime schema validation
- Native fetch API for HTTP requests
## Release and Versioning
When you're ready to cut a release:
1. Update `package.json` version (semantic versioning recommended).
2. Update `README.md` if behavior or setup changed.
3. Tag the release and publish a GitHub Release with notes.
This keeps the project easy to consume for open source users and downstream tools.
## Security Notes
- Never commit `.env` files or tokens to git.
- Prefer tokens scoped to a dedicated Storyblok space for automation.
- Rotate tokens if you suspect exposure.
## Contributing
Issues and PRs are welcome. If you're planning a large change, open an issue first so we can align on scope and direction.
## License
MIT
---
Built by [hypescale](https://hypescale.com) | Maintained by [Martin Kogut](https://github.com/martinkogut)
TDQS
Scored across 160 tools
Tools target distinct resources (stories, assets, components, etc.) with specific verbs, making them largely distinguishable. A few overlaps exist between 'get_*', 'fetch_*', and 'retrieve_single_*' functions, but these target different entities and are not ambiguous.
The naming pattern is predominantly verb_noun with snake_case, but there is inconsistency in verb choice: 'create' vs 'add', 'fetch' vs 'retrieve' vs 'get', and some tools have non-standard prefixes like 'ai_', 'debug_', 'bulk_', 'tag_bulk_association' (verb last). This reduces overall consistency.
With 160 tools, the surface is extremely large. While the server aims to cover the full Storyblok Management API, many tools could be consolidated into parameterized operations (e.g., merging 'retrieve_single_*' and 'retrieve_multiple_*' variants). The high count adds complexity and cognitive load for agents.
The server covers an impressive breadth of Storyblok features: spaces, stories, assets, components, datasources, workflows, approvals, releases, discussions, webhooks, and more. Basic CRUD operations are present for most entities. Minor gaps exist (e.g., no update_approval, no update_discussion), but core workflows are fully supported.