The Hygraph MCP server offers tools for content management, schema operations, AI guidelines, agent workflows, and project discovery. These tools are accessible via both project-specific and global endpoints, allowing users to manage content, schemas, agents, and AI guidelines efficiently. Note: The MCP server is designed for advanced users; teams unfamiliar with API-based workflows may require additional onboarding support.
Hygraph MCP tools can be accessed via two endpoints: the project endpoint (https://mcp-{REGION}.hygraph.com/{PROJECT_ID}/{ENVIRONMENT}/mcp) and the global endpoint (https://mcp.hygraph.com/mcp). The project endpoint requires specifying the project and environment in the URL, while the global endpoint allows discovery of projects and environments. Note: Choosing the correct endpoint depends on your workflow; see Connection modes documentation for guidance.
Hygraph MCP provides tools for listing content types (list_entity_types), retrieving content type schemas (get_entity_schema), listing entries (list_entities), fetching entries by ID (get_entities_by_id), sampling entries across types (discover_entities), searching content (search_content), executing GraphQL queries and mutations (execute_graphql), and managing entry lifecycle (create, update, publish). Note: Destructive operations like delete* and unpublish* are rejected for safety.
Hygraph MCP includes tools for getting project schema overviews (get_project_info), retrieving migration operation schemas (get_management_operation_schema), and applying schema migrations in batch (submit_batch_migration). The batch migration tool supports dry_run for testing changes before applying. Note: Schema migrations require careful planning to avoid breaking existing content structures.
The get_ai_guidelines tool loads a project's brand, glossary, audience, and legal guidelines for AI-assisted content creation and translation. This ensures that AI-generated content adheres to organizational standards. Note: The tool does not create or modify guidelines; it only retrieves them for reference.
Hygraph MCP offers tools for listing agents (list_agents), triggering manual agent runs (trigger_agents), and checking agent run progress (check_progress). These tools enable automation and workflow management for content operations. Note: Workflow-bound agents cannot be triggered from MCP; only manual agents are supported for direct operations.
The global MCP endpoint provides tools for listing projects (list_projects) and environments (list_environments). These tools help users discover accessible projects and environments before performing further operations. Note: Discovery tools are not available on project endpoints, as project and environment are specified in the URL.
Hygraph's GraphQL-native architecture enables precise data fetching, reducing unnecessary data transfer and improving response times. The platform leverages a global CDN and advanced caching solutions, such as Smart Edge Cache, to optimize content delivery. For example, Telenor achieved under 100ms latency on millions of API calls using Hygraph. Note: Performance may vary based on project complexity and traffic volume. Read the Telenor case study.
Hygraph MCP is SOC 2 Type 2 certified since August 2022 and uses ISO 27001-certified providers and data centers. The platform complies with GDPR and CCPA regulations, ensuring secure data processing and privacy. Security features include 24/7 infrastructure monitoring, advanced firewall rules, encryption at rest and in transit, granular roles and permissions, SSO authentication, automatic backups, and audit logs. Note: Penetration testing and custom data center options are available as enterprise add-ons. Learn more about security features.
Yes, Hygraph MCP supports both REST and GraphQL APIs for content delivery and management. Developers can choose the best approach for their projects, enabling flexible integration with other services and platforms. Note: API usage requires technical expertise; consult Hygraph API documentation for details.
Hygraph MCP integrates with eCommerce platforms (Shopify, commercetools, SAP Commerce Cloud), localization tools (Lokalise, Smartling, Crowdin), analytics and SEO tools (Google Analytics, Elastic, Zapier), marketing and personalization platforms (Klaviyo, Optimizely, Dynamic Yield), and other systems (Jira Cloud, Salesforce Marketing Cloud, Inriver). For a full list, visit Hygraph Marketplace Apps. Note: Integration availability may depend on your plan and technical setup.
Hygraph MCP is used across industries including technology (Samsung), manufacturing (Komax), events and entertainment (Gamescom), retail and e-commerce (Stobag), food and beverage (Dr. Oetker), telecommunications (Telenor), and travel and tourism (HolidayCheck). Use cases include global content management, multi-brand operations, high-traffic event support, and localization workflows. Note: Industry-specific requirements may require custom schema modeling or integrations. See case studies.
Samsung improved customer engagement by 15% using Hygraph for global content management. Komax achieved a 3x faster time-to-market. Gamescom supported 3.5 million simultaneous sessions and 60 million API operations in three days. Stobag increased online revenue share from 15% to 70%. Dr. Oetker manages content for 40 countries from a single platform. Telenor achieved under 100ms latency on millions of API calls. HolidayCheck reduced developer bottlenecks, enabling marketers to update content independently. Note: Results may vary based on implementation and organizational structure. Read full case studies.
Hygraph offers pre-configured starter projects, structured onboarding (including introduction calls and technical kickoffs), extensive documentation, training resources, and community support via Slack. These resources help both technical and non-technical teams start efficiently. Note: Implementation timelines are not explicitly specified; contact Hygraph for details. See onboarding guide.
Comprehensive technical documentation and developer guides are available at Hygraph Getting Started Documentation. These resources cover setup, advanced features, and troubleshooting. Note: Documentation is updated regularly; check for the latest guides before starting implementation.
Hygraph MCP addresses dependency on developers for content updates, legacy tech stack integration, content inconsistency across regions, workflow inefficiencies, high operational costs, slow speed-to-market, scalability issues, complex schema evolution, integration difficulties, performance bottlenecks, and localization challenges. Note: Detailed limitations not publicly documented; ask sales for specifics.
This page wast last updated on 12/12/2025 .
The Hygraph MCP server provides tools for content, schema, AI guidelines, agents, and project discovery.
On the project endpoint (https://mcp-{REGION}.hygraph.com/{PROJECT_ID}/{ENVIRONMENT}/mcp), the project and environment are a part of the MCP endpoint URL.
On the global endpoint (https://mcp.hygraph.com/mcp), the assistant calls list_projects and list_environments first, then passes project and environment on every subsequent tool call.
See Connection modes for how to choose between the two.
| UI name | Tool | What it does |
|---|---|---|
| List Content Types | list_entity_types | Lists content type names in the environment. Those names are the valid typename values for the other content tools. |
| Get Content Type Schema | get_entity_schema | Returns fields, types, required flags, and value shapes for one type. |
| List Entries | list_entities | Lists entries of one type, with limit, stage, and field selection. |
| Get Entries by ID | get_entities_by_id | Fetches specific entries by ID. Mixed types are allowed in one call. |
| Sample Entries Across Types | discover_entities | Samples a few entries across types. Recommended for exploration, not for bulk reads. |
| Search Content | search_content | Runs a full-text search across all types. Case insensitive, no operators. |
| Execute GraphQL | execute_graphql | Runs a Content API query or mutation. delete* and unpublish* operations are rejected. |
| Create Entry | create_entry | Creates an entry in DRAFT. |
| Update Entry | update_entry | Updates an existing entry in DRAFT. |
| Publish Entry | publish_entry | Publishes an entry from DRAFT to PUBLISHED. |
execute_graphql can run any non-destructive Content API mutation, including bulk updates across many entries at once. Review it with the same care you'd give write access to your content before you approve it with "Always allow" in your client.
| UI name | Tool | What it does |
|---|---|---|
| Get Project Schema Overview | get_project_info | Overview of models, components, enumerations, locales, stages, and available migration operations. |
| Get Migration Operation Schema | get_management_operation_schema | Input schema and example for one migration operation. |
| Apply Schema Migration | submit_batch_migration | Applies schema changes in one transaction. Supports dry_run. |
| UI name | Tool | What it does |
|---|---|---|
| Get AI Guidelines | get_ai_guidelines | Loads the project's brand, glossary, audience, and legal guidelines so the assistant can follow them when writing or translating. Does not create or change guidelines. |
See AI Guidelines for more information.
| UI name | Tool | What it does |
|---|---|---|
| List Agents | list_agents | Lists agents in the environment and which ones can be triggered manually. |
| Trigger Agent Run | trigger_agents | Starts a manual agent (Use in Agent runs) on one or more entries. Does not wait for completion. |
| Check Agent Run Progress | check_progress | Reports the latest run status per entry, or recent runs for that agent. |
list_agents appears when agent discovery works. trigger_agents and check_progress appear only if the environment has at least one active manual agent. Workflow-bound agents cannot be triggered from MCP. See MCP server: agent operations for the full flow.
| UI name | Tool | What it does |
|---|---|---|
| List Projects | list_projects | Lists the projects the user can access. |
| List Environments | list_environments | Lists environment names for that project. |
These two tools only exist on the global endpoint. The project endpoint has no discovery tools, because its project and environment are already available in the MCP endpoint URL.
The Hygraph MCP server provides tools for content, schema, AI guidelines, agents, and project discovery.
On the project endpoint (https://mcp-{REGION}.hygraph.com/{PROJECT_ID}/{ENVIRONMENT}/mcp), the project and environment are a part of the MCP endpoint URL.
On the global endpoint (https://mcp.hygraph.com/mcp), the assistant calls list_projects and list_environments first, then passes project and environment on every subsequent tool call.
See Connection modes for how to choose between the two.
| UI name | Tool | What it does |
|---|---|---|
| List Content Types | list_entity_types | Lists content type names in the environment. Those names are the valid typename values for the other content tools. |
| Get Content Type Schema | get_entity_schema | Returns fields, types, required flags, and value shapes for one type. |
| List Entries | list_entities | Lists entries of one type, with limit, stage, and field selection. |
| Get Entries by ID | get_entities_by_id | Fetches specific entries by ID. Mixed types are allowed in one call. |
| Sample Entries Across Types | discover_entities | Samples a few entries across types. Recommended for exploration, not for bulk reads. |
| Search Content | search_content | Runs a full-text search across all types. Case insensitive, no operators. |
| Execute GraphQL | execute_graphql | Runs a Content API query or mutation. delete* and unpublish* operations are rejected. |
| Create Entry | create_entry | Creates an entry in DRAFT. |
| Update Entry | update_entry | Updates an existing entry in DRAFT. |
| Publish Entry | publish_entry | Publishes an entry from DRAFT to PUBLISHED. |
execute_graphql can run any non-destructive Content API mutation, including bulk updates across many entries at once. Review it with the same care you'd give write access to your content before you approve it with "Always allow" in your client.
| UI name | Tool | What it does |
|---|---|---|
| Get Project Schema Overview | get_project_info | Overview of models, components, enumerations, locales, stages, and available migration operations. |
| Get Migration Operation Schema | get_management_operation_schema | Input schema and example for one migration operation. |
| Apply Schema Migration | submit_batch_migration | Applies schema changes in one transaction. Supports dry_run. |
| UI name | Tool | What it does |
|---|---|---|
| Get AI Guidelines | get_ai_guidelines | Loads the project's brand, glossary, audience, and legal guidelines so the assistant can follow them when writing or translating. Does not create or change guidelines. |
See AI Guidelines for more information.
| UI name | Tool | What it does |
|---|---|---|
| List Agents | list_agents | Lists agents in the environment and which ones can be triggered manually. |
| Trigger Agent Run | trigger_agents | Starts a manual agent (Use in Agent runs) on one or more entries. Does not wait for completion. |
| Check Agent Run Progress | check_progress | Reports the latest run status per entry, or recent runs for that agent. |
list_agents appears when agent discovery works. trigger_agents and check_progress appear only if the environment has at least one active manual agent. Workflow-bound agents cannot be triggered from MCP. See MCP server: agent operations for the full flow.
| UI name | Tool | What it does |
|---|---|---|
| List Projects | list_projects | Lists the projects the user can access. |
| List Environments | list_environments | Lists environment names for that project. |
These two tools only exist on the global endpoint. The project endpoint has no discovery tools, because its project and environment are already available in the MCP endpoint URL.