Fireflies MCP Service Adapter

This module provides a programmatic interface (MCP Server) for interacting with the Fireflies.ai Application Programming Interface (API), focusing on meeting data lifecycle management, specifically retrieval, content indexing, and abstract generation.

Core Capabilities

• Record Index Access: Secure fetching and browsing of stored meeting transcripts.
• Data Exploration: Application of sophisticated search predicates to locate specific conversation segments or documents.
• Knowledge Distillation: Automated creation of condensed summaries from lengthy conversational data, available in multiple structural styles.

Exposed Functions

1. fireflies_get_transcripts
2. Purpose: Fetches a collection of meeting artifacts, supporting temporal and count constraints.
3. Parameters:
   • limit (Optional Integer): Caps the cardinality of the returned result set.
   • from_date (Optional Text): ISO 8601 compliant starting boundary for the search period (e.g., YYYY-MM-DD).
   • to_date (Optional Text): ISO 8601 compliant ending boundary for the search period (e.g., YYYY-MM-DD).

4. Output: A serialized array containing foundational metadata for each retrieved transcript.

5. fireflies_get_transcript_details

6. Purpose: Retrieves the complete, granular content and associated metadata for a designated conversation.
7. Parameters:
   • transcript_id (Required Text): The unique identifier string pointing to the target transcription.

8. Output: A detailed object encompassing speaker identification, verbatim text, and auxiliary contextual information.

9. fireflies_search_transcripts

10. Purpose: Executes full-text indexing searches across the corpus based on user-defined strings.
11. Parameters:
    • query (Required Text): The string or phrase to locate within the transcript contents.
    • limit (Optional Integer): Maximum volume of matching transcripts to be returned.

12. Output: A subset of transcript metadata objects whose content satisfies the search predicate.

13. fireflies_generate_summary

14. Purpose: Invokes the summarization engine for a specific transcript.
15. Parameters:
    • transcript_id (Required Text): The unique identifier of the source meeting record.
    • format (Optional Text): Dictates the resultant summary structure ('bullet_points' or 'paragraph').
16. Output: The synthesized, condensed textual representation of the meeting's substance.

Deployment Configuration (Claude Desktop)

To integrate this service within the Claude Desktop environment, the following structure must be incorporated into your claude_desktop_config.json file. Ensure the required API credential is substituted for <YOUR_API_KEY>.

{ "mcpServers": { "fireflies": { "command": "npx", "args": [ "-y", "@props-labs/mcp/fireflies" ], "env": { "FIREFLIES_API_KEY": "" } } } }

Credential Acquisition

Authentication requires a valid API key obtained from the Fireflies.ai portal: 1. Access the Fireflies.ai administrative interface. 2. Navigate to the configuration section dedicated to APIs (Settings > API). 3. Initiate the generation of a novel access token. 4. Securely retain the generated key for configuration purposes.

Initializing the Service Runtime

1. Obtain the source code repository.
2. Resolve required software packages:

bash npm install

Alternatively:

pnpm install

1. Compile the project assets:

bash npm run build

Alternatively:

pnpm build

Execution Command

To initiate the server, specify the environment variable containing the authentication token:

bash FIREFLIES_API_KEY=your_secret_token npm start

Or via pnpm:

FIREFLIES_API_KEY=your_secret_token pnpm start

Legal Disclosures

This server adapter operates under the stipulations of the MIT License. Review the accompanying LICENSE documentation within the project structure for comprehensive rights concerning utilization, modification, and redistribution.