Builder MCP. Write code to manage your CMS data.
Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Builder MCP Server gives your AI agent direct control over your Builder.io headless CMS. You can programmatically list models, fetch their exact schemas, count content entities, and generate or update visual blocks—all through natural conversation in your code editor.
What your AI agents can do
Count model entities
Quickly counts the number of active content entries stored within a specific model.
Create visual block
Generates and inserts new visual content blocks or full records into a specified Builder model.
Get media file
Retrieves metadata, including URLs, for an uploaded media asset in your Builder space.
List all defined data models in your space, check their current schema structure, and find specific models by name.
Generate new visual content blocks or update existing ones using the model's required JSON format.
Pull a list of all entries for a given model, or retrieve one specific document by title and query string.
Permanently delete content blocks or components that are deprecated or orphaned in your system.
Get metadata, including URLs and details, for any media file attached to a block or model.
Ask AI about this MCP
Supported MCP Clients
Waiting for input…
Builder MCP Server: 10 Tools for Headless CMS Ops
Use these tools to read model schemas, generate content entries, count records, or delete components in Builder.io—all via natural language prompts.
019d7565count model entities
Quickly counts the number of active content entries stored within a specific model.
019d7565create visual block
Generates and inserts new visual content blocks or full records into a specified Builder model.
019d7565get media file
Retrieves metadata, including URLs, for an uploaded media asset in your Builder space.
019d7565get model schema
Provides the exact field structure and JSON definition required by a single content model.
019d7565get single content
Fetches one specific content document from a model using targeted query matching (e.g., finding 'Home').
019d7565list builder models
Lists every defined data model and schema available in your Builder workspace.
019d7565list model content
Retrieves a complete list of content entries for an entire model, useful for bulk operations.
019d7565search active models
Finds and lists Builder models that match a specific name or substring criteria.
019d7565update visual block
Modifies the data fields of an existing content block or visual document within a model.
019d7565wipe visual block
Permanently deletes a specific, targeted content entry from Builder.io.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Builder, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Builder MCP Server: Direct Control Over Your CMS Content.
You're tired of switching between your code editor and the Builder UI just to manage content? This server gives your AI agent direct, programmatic control over every piece of content in your headless CMS. You don't need any manual steps; your agent handles listing models, checking schemas, counting records, and generating or modifying visual blocks—all right within your natural conversation flow.
Managing Your Models and Schemas
You can start by having the agent list every single data model defined in your Builder workspace using list_builder_models. If you know a part of the name, you don't have to read everything; the system lets you narrow it down with search_active_models, finding models that match specific names or substrings.
Once you pick a model, you can use get_model_schema to pull the exact field structure and JSON definition required for that content type. This tells your agent precisely what data format you need before it writes anything.
Reading and Fetching Content Data
When you need content, you've got options. To get a full picture of everything in one model, use list_model_content to pull a complete list of entries for that entire model—perfect for bulk operations or auditing. If you only need one specific document, the agent can fetch it using get_single_content, which requires targeted query matching (like finding the 'Home' page).
You can also quickly check how many active content entries exist in any given model by calling count_model_entities. For handling media attachments, you just use get_media_file to retrieve metadata, including URLs and details, for any uploaded asset attached to a block or model.
Creating and Updating Blocks
The agent can build content from scratch. You can generate and insert new visual content blocks or full records into a specified Builder model using create_visual_block. Need to tweak an existing piece? The system lets you modify the data fields of any current block or visual document within its model structure via update_visual_block.
This means your agent doesn't just read; it makes changes, keeping your content fresh without ever opening the dashboard.
Auditing and Cleaning Up
When a page or component is deprecated, you need to get rid of it. The server lets you permanently delete a specific, targeted content entry from Builder.io using wipe_visual_block. This gives you full control over cleaning out old or orphaned components in your system.
How Builder MCP Works
- 1 Subscribe to the server and provide your Builder Public and Private API Keys.
- 2 Start talking to the agent. Ask it to list models, check schemas (e.g.,
list_builder_models), or find a specific content ID. - 3 The agent calls the necessary tools, reads the schema data, and executes the operation—all displayed in your chat window.
The bottom line is you get to treat your entire visual CMS like another API endpoint inside your code editor.
Who Is Builder MCP For?
This server is for developers and content ops who hate context switching. If you're tired of opening the Builder UI just to check a schema or push 50 translated records, this is for you. It turns CMS management into another function call.
You use get_model_schema to validate the exact data structure of a model before writing fetch logic in React or Vue.
You automate bulk content tasks, like running create_visual_block 50 times for localized copy based on an existing template.
You run cleanup scripts using wipe_visual_block to delete orphaned test data or deprecated components without manual UI steps.
What Changes When You Connect
- Schema Validation: Use
get_model_schemafirst. You always know the exact field types and JSON bounds before attempting a write, eliminating runtime errors. - Bulk Content Generation: Instead of manual uploads, use
create_visual_blockto generate dozens of content variations (e.g., translated landing page copies) with one agent call. - Precise Targeting: The combination of
get_single_contentand query strings means you pull exactly the right document ID without guessing or searching through massive lists. - System Auditing: Use
count_model_entitiesto get a real-time count of records. This is faster than clicking into multiple model dashboards just for a number. - Lifecycle Management: When content becomes obsolete, use
wipe_visual_block. You delete data programmatically and confirm the removal instantly.
Real-World Use Cases
The Weekly SEO Audit
A team member needs to check every 'blog-post' model for a specific tag. They ask their agent to run list_model_content on the blog-post model, then filter the results by query string, and finally use get_media_file on any images found that lack alt text.
The A/B Test Rollout
You need to test two versions of a hero block. You ask the agent to first run list_builder_models to confirm the 'hero' model, then use get_model_schema to get the required fields, and finally execute create_visual_block twice with slightly different content payloads.
Cleaning Up Stale Data
The project is sunsetting an old feature. Instead of manually finding all related pages, a developer tells the agent to run search_active_models for 'old-feature' and then use wipe_visual_block on every document found.
Checking Data Consistency
Before deploying a new page template, you need to know if the model is set up correctly. You first run list_builder_models, pick the target model, and then call get_model_schema to ensure all required fields are present.
The Tradeoffs
Trying to guess content IDs
The developer thinks they know a document ID off the top of their head and tries to retrieve it. The agent returns an error because the ID is wrong or deprecated.
→
Start by running list_model_content on the correct model, then use get_single_content with reliable query parameters like title or slug. Never guess.
Overwriting data without checking
A developer uses update_visual_block but forgets to pass the entire payload, only sending a partial update. The CMS rejects it because required fields are missing.
→
Always run get_model_schema first. This guarantees you have all necessary keys and data types needed for a successful update.
Running cleanup scripts blindly
A DevOps engineer runs wipe_visual_block on a model without checking the scope, accidentally deleting live production content.
→
Always run list_model_content or count_model_entities first. Verify the count and confirm the target IDs before executing any deletion tool.
When It Fits, When It Doesn't
Use this server if your job involves managing structured, headless content—i.e., you need to programmatically read schemas, create records, or manage assets within a Builder-like CMS. You should use it when the data source is already defined by a clear schema (which Builder provides). Don't use it if you just need general file storage management; for that, look into generic object storage APIs. If your process requires complex human decisioning between content types or manual approvals before publishing, this server only handles the mechanics of the write/read cycle. You still need a workflow tool (like Zapier or dedicated CI/CD) to manage the 'when' and 'why,' but this server lets you handle the 'how.'
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Builder. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Checking content status shouldn't mean leaving your terminal.
Today, if you need to know what fields a component needs—say, `cta_link` and `title`—you have to jump into the Builder UI. You click around until you find the documentation or model definition page, copy the schema details, and paste them back into your terminal session just to verify the data types.
With this MCP server, you ask the agent: 'What is the schema for the hero section?' It runs `get_model_schema` instantly. The full field list, required status, and JSON bounds appear right in your chat window. You get the structure without leaving your code editor.
The `list_builder_models` tool gives you instant visibility into everything.
Before writing any single line of fetch logic, you have to know what models exist. Do they even have a 'page' model? Is the blog data stored in 'posts' or 'blog-article'? Right now, figuring out your CMS architecture is a manual process of clicking through menus and guessing endpoints.
Now, just ask the agent to run `list_builder_models`. It dumps every single defined model name right back to you. This makes building integrations faster; you know exactly what data containers are available.
Common Questions About Builder MCP
How do I check the structure of a content block using get_model_schema? +
You run get_model_schema and provide the model name. The agent returns the exact field list, including required fields and their data types (Text, URL, File/URL). This tells you exactly what JSON to send.
Can I find all content entries for a specific model? Use list_model_content. +
Yes. list_model_content fetches the complete roster of every item in that model, giving you IDs and titles. This is your starting point if you need to process many records.
How do I delete content using wipe_visual_block? +
You must provide a specific target ID or unique query string. The agent runs wipe_visual_block and confirms the permanent deletion of that single document from your Builder space.
What if I only want one page, not all content? Should I use get_single_content? +
Yes. If you know a unique identifier or title (like 'About Us'), get_single_content uses query parameters to pull just that document. It's much more precise than listing the whole model.
When I run `list_builder_models`, what credentials do I need to provide for authentication? +
You must use the Builder Public and Private API Key pair. These keys authenticate your AI client against the Builder service, allowing it access to list all defined models. This ensures that only authorized connections can map out your CMS structure.
If I need to know the total item count for a model, is `count_model_entities` efficient? +
Yes, count_model_entities performs a direct query count, which is far faster than fetching and counting individual records. It returns a single integer representing the current number of active items in that specific data model.
How do I pull asset details from an image using `get_media_file`? +
The tool retrieves detailed metadata, including the direct URL and dimensions for any uploaded asset. You get a structured object containing all necessary information to display or process the media on your site.
If I use `create_visual_block`, what happens if my content data doesn't match the schema? +
The tool will return an explicit error message detailing which fields are missing or malformed. Always check the returned JSON object for validation failures; this tells you exactly how to fix your payload before attempting a second creation attempt.
Can my AI read the strict structure of a custom model before creating a block? +
Yes! We highly recommend asking your agent to use get_model_schema first. By retrieving the exact JSON fields (Strings, Lists, Booleans), the AI naturally learns the shape of your Page or HeroContent model and formats its create_visual_block mutation flawlessly avoiding validation errors.
How can I count how many blogs or pages are in my Builder space? +
You can use the native count_model_entities tool. By querying a model like blog-post, the agent accesses the Builder backend with limit=1 exclusively to extract the metadata counts without downloading heavyweight content arrays—highly efficient for broad analytics.
Can the agent wipe an accidentally created page completely? +
Yes. If an error occurs or a staging draft needs removal, the agent can call wipe_visual_block. Supply the targeted entity ID, and the HTTP DELETE command is irreversibly broadcasted to the Builder environment. Always review IDs manually.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Notion Calendar (formerly Cron)
Manage scheduling via Notion Calendar — create events, track team availability, and manage scheduling links directly from any AI agent.
GitHub Alternative
Manage repositories, issues, pull requests and workflows via GitHub — search code, audit commits and track CI/CD from any AI agent.
HashiCorp Nomad
Manage workloads and orchestration via Nomad — track jobs, nodes, and deployments directly from your AI agent.
You might also like
Fountain
Automate high-volume hiring, manage applicant funnels, and track hiring goals via AI agents with Fountain.
Untappd
Track beers, breweries, and venues. Access check-in feeds, beer metadata, and manage your social drinking profile directly from your AI agent.
Anyscale
Orchestrate your Anyscale infrastructure — manage LLM queries, vectors, services, and cluster batch jobs directly from your AI agent.