General Questions
Q: What is the Gracenote MCP server?
A: Gracenote is launching a new technology and tool that will make Large Language Models (LLMs) smarter and more accurate using Gracenote metadata. For the first time, Gracenote is making its normalized and structured metadata searchable and queryable via a Model Context Protocol (MCP) server for AI use cases. With this new offering, TV and video platforms and apps can employ LLMs to power entertainment search and discovery use cases, verified and enriched by Gracenote data.
With each real-time query, LLMs will be able to verify their contextually relevant responses against Gracenote data. This new product handles real-time queries and retrieves intelligence from Gracenote's entertainment knowledge base made up of Gracenote IDs (TMS), factual data (title, year, genre) and descriptive data (description). It also has the capability to deliver imagery URLs, content, availability information, and related deep links.
Q: What does MCP mean?
A: MCP (Model Context Protocol) is an open-source standard that helps LLMs and other AI models give better, more trustworthy answers. MCP Connects LLMs to external tools: MCP acts as a universal adapter, letting LLMs access information from external tools, systems and sources (such as websites or databases). MCP Grounds LLM responses: Before the LLM generates its response, it uses MCP in real-time to verify that the answer is up to date, accurate, and if needed enriches it with additional information.
Q: What are the biggest key values of the Video MCP Server?
A: With the Gracenote Video MCP Server, customers can:
- Verify and enrich LLM results in real-time (an exercise known as grounding) with Gracenote's trusted, industry leading entertainment metadata, imagery and IDs:
- Reduces errors and hallucinations
- Makes TMSIDs and images accessible to LLMs
- Provides up-to-date availability information
- Leverages LLM semantic search and powerful results ordering
- Rapidly integrate a Gracenote maintained solution, reducing data engineering and ingestion costs
- Power engaging, content-first UX with advanced conversational search, tailored recommendations, compelling discovery paths and universal 'where to watch' details
- Enable catalog enrichment and data harmonization across disparate entertainment data sources, leveraging Gracenote IDs for content identification
- Support both user-facing applications (chatbots, search) and machine-to-machine integrations (batch processing, scheduled jobs)
Q: How does the MCP Server differ from OnConnect and OnAPI?
A: Gracenote OnConnect and On API deliver specific data sets, while the Video MCP Server is geared towards LLMs and makes Gracenote data sets and logic accessible and understandable to LLMs. The Video MCP Server enables querying to look up content by title, year, genre, etc. as well as where to watch programming.
Access & Onboarding
Q: How do I get access to the MCP Server?
A: Contact your Gracenote representative to discuss access and onboarding. Access includes:
- MCP Server credentials (Cognito username/password, client ID/secret)
- Customer dashboard access
- Unique MediaCloud subdomain for images
- Documentation and support resources
Q: What credentials do I need?
A: You will receive:
- MCP Server credentials (username, password, client ID, client secret)
- MCP Server host information
- Customer dashboard login credentials
- MediaCloud subdomain for image URLs (for production deployment)
Q: How do I track my usage?
A: Use the Customer Dashboard to:
- View usage by day, week, month, year
- See usage breakdown by tool
- View specific call inputs/outputs
- Export usage data as CSV
Data & Coverage
Q: What content types are supported?
A: The MCP Server returns:
- Feature Films
- TV Movies
- TV Series
- TV Episodes
Excluded: Sports events, previews, bonus content, non-TV/movie content
Q: Is sports data available?
A: Sports data will be available via a separate Sports MCP Server in the future.
Technical Questions
Q: What MCP tools are available?
A: The MCP Server provides five tools that your LLM uses automatically:
- Entity Resolution: Find and disambiguate movies, TV shows, and episodes
- Streaming Availability: Determine where content is available to watch
- Root Tool: Find alternate language versions of titles
- TMSID Tool: Get detailed metadata for specific titles
- Image Tool: Access all available images for titles
Your LLM selects and invokes these tools based on user queries. See Tool Reference for details.
Q: Can I call the tools directly?
A: In general, you don't call these tools directly - your LLM does. You can call these tools using MCP Inspector for testing purposes. See Option 2 in the Quick Start (CLI/Proxy)
Q: How long do authentication tokens last?
A: Cognito tokens expire after 1 hour. Implement token refresh logic for long-running processes. See Connecting to the MCP Server for examples.
Q: What happens if I exceed input limits?
A: The server returns human-readable error messages indicating proper usage. For example:
- Entity Resolution accepts maximum 3 cast members
- Exceeding limits returns an error with guidance
LLMs can self-correct and retry with proper input.
Support & Troubleshooting
Q: How do I get support?
A: Multiple support channels available:
- Submit support tickets via Customer Dashboard
- Contact your Gracenote representative
Q: What should I do if authentication fails?
A: Common authentication errors:
| Error | Cause | Solution |
|---|---|---|
| NotAuthorizedException | Invalid username/password | Verify credentials |
| UserNotConfirmedException | User not confirmed | Contact Gracenote to confirm user |
| 401 Unauthorized | Invalid/expired token | Get fresh token |
| invalid_token | Server rejects token | Contact Gracenote support |
See Connecting to the MCP Server for authentication code examples and troubleshooting.
Q: What if I get no results from Entity Resolution?
A: Check:
- Language and country parameters are correct
- Title spelling is accurate (tool handles partial matches)
- Cast member count doesn't exceed 3
- Content type is supported (movies, series, episodes only)
Q: How do I report a bug or request a feature?
A: Submit via Customer Dashboard or contact your Gracenote representative.
Q: Where can I find code examples?
A: See documentation for complete examples:
- Connecting to MCP - MCP client setup with authentication
- Catalog Harmonization - Batch processing, ETL, scheduled jobs
- Search & Discovery - User-facing search with Streamlit
- Personalized Recommendations - CLI and REST API recommendation engines
Demo & Testing
Q: Is there a demo available?
A: Yes. See Gracenote MCP Demo for access instructions and example queries.