How to Connect SayItVid to Claude, Cursor & ChatGPT via MCP: Real-Time Pronunciation & Video Data

Introduced by Anthropic as an open industry standard, the Model Context Protocol (MCP) enables AI systems to connect to external databases, tools, and real-time APIs using standardized JSON-RPC protocols. Rather than relying on static training weights that may be outdated or inaccurate, an MCP-connected agent can invoke specialized tools dynamically during a chat or coding session.

Ask standard GPT-4 or Claude 3.5 to transcribe a technical term or dialect variant in IPA, and you will often receive subtle errors: incorrect stress markers, missing glottal stops, or invented vowel symbols. Furthermore, text-only models cannot provide authentic audio clips or demonstrate real human speakers articulating words in genuine professional contexts.

By grounding your assistant with SayItVid MCP, every answer is backed by curated phonetic dictionaries and timestamped YouTube video moments.

The SayItVid MCP server exposes four dedicated production tools:

1. 1search_pronunciation: Search hundreds of thousands of conversational video clips by keyword, part of speech, or regional accent (US, UK, AUS).
2. 2get_word_details: Retrieve authoritative IPA transcriptions, grammatical categories, definitions, and syllable breakdowns.
3. 3get_syllables_and_stress: Analyze exact syllable counts, primary and secondary stress positions, and haiku line meter.
4. 4compare_accents: Query side-by-side dialect comparisons between American, British, and Australian English.

Cursor IDE features built-in native support for remote SSE MCP servers without needing local Node.js or Python packages:

1. 1Open Cursor Settings (Cmd+, on macOS or Ctrl+, on Windows/Linux).
2. 2Navigate to Features $\rightarrow$ MCP.
3. 3Click + Add New MCP Server.
4. 4Enter the configuration:

• Name: sayitvid
• Type: sse
• URL: https://sayitvid.com/api/mcp

1. 1Click Add. The status indicator will switch to green, confirming that all 4 tools are loaded.

To connect Claude Desktop to SayItVid's hosted HTTP endpoint, use the official mcp-remote bridge:

1. 1Open claude_desktop_config.json:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

1. 1Add the server under mcpServers:

json { "mcpServers": { "sayitvid": { "command": "npx", "args": ["-y", "mcp-remote@latest", "https://sayitvid.com/api/mcp"] } } }

1. 1Restart Claude Desktop. The hammer icon in the message bar will display the active SayItVid tools.

If you use modern developer agent harnesses like Opencode or Google Antigravity, add the remote MCP entry directly to your project or user configuration file (opencode.json):

json { "mcp": { "sayitvid": { "type": "remote", "url": "https://sayitvid.com/api/mcp" } } }

Your autonomous coding agents can now verify pronunciation and syllable counts when generating language-learning exercises or writing documentation.

Custom Python scripts, LangChain agents, or ChatGPT Custom Actions can query SayItVid directly using standard JSON-RPC 2.0 over HTTP POST:

bash curl -X POST https://sayitvid.com/api/mcp \ -H "Content-Type: application/json" \ -d '{ "jsonrpc": "2.0", "id": "1", "method": "tools/call", "params": { "name": "get_word_details", "arguments": { "word": "ephemeral" } } }'

Developers and educators are utilizing SayItVid MCP for several high-impact use cases:

• AI ESL Tutoring Agents: When explaining difficult vocabulary, the AI calls search_pronunciation to supply the student with authentic video timestamp clips rather than dry text explanations.
• Technical Documentation Pronunciation Badges: Automatically audit open-source library documentation to inject correct IPA and spoken audio links for tricky package names.
• Automated Poetry and Lyric Generators: Verify that generated sonnets or haikus strictly conform to phonetic syllable counts using get_syllables_and_stress.