Fireflies Tools Reference

Overview

When you connect Fireflies to your MCP server, you get 3 additional tools that provide comprehensive access to your meeting transcripts. These tools work together to list transcripts, search across all meetings, and retrieve detailed conversation data.

Available Tools

Typical Workflows

Search → Transcript Retrieval

Example natural language command:

"Search my Fireflies transcripts for mentions of 'product roadmap',
 then create a mind from the most recent match"

Discovery → Analysis Workflow

Example natural language command:

"Show me my team meetings from last month, then create
 minds for each team member based on their participation"

Tool Details

fireflies_list_transcripts

Purpose: List meeting transcripts with metadata and date filtering

Key Features:

• Date range filtering (from/to dates)
• Pagination support for large result sets
• Returns transcript IDs, titles, participants, duration, date
• Quick overview of available meetings

When to use:

• Finding specific transcripts by date
• Discovering available meetings before retrieval
• Building transcript inventories
• Browsing recent discussions

View full documentation →

fireflies_get_transcript

Purpose: Retrieve detailed transcripts with speaker analytics

Key Features:

• Complete conversation text
• Speaker names (emails inferred via name matching)
• Timestamp information
• Structured by speaker turns
• Speaker analytics (talk time, word count, speaking patterns)

When to use:

• Creating minds from meeting conversations
• Analyzing speaker participation
• Extracting individual speaker segments
• Training AI on actual conversations

View full documentation →

fireflies_search_transcripts

Purpose: Search across all transcripts by keyword

Key Features:

• Keyword search across full transcript text
• Date range filtering combined with keyword search
• Find specific topics or phrases
• Returns matching transcripts with relevance

When to use:

• Finding discussions about specific topics
• Locating mentions of products, features, or customers
• Discovering relevant conversations across time
• Building topic-specific mind training sets

View full documentation →

Integration with Mind Reasoner

Creating Minds from Fireflies Transcripts

The most common workflow combines Fireflies tools with Mind Reasoner tools:

1. Discovery: Use fireflies_search_transcripts or fireflies_list_transcripts to find relevant meetings
2. Retrieval: Use fireflies_get_transcript to get the full conversation
3. Mind Creation: Use Mind Reasoner create_mind to create a digital entity
4. Upload: Upload the Fireflies transcript as training data
5. Training: Create snapshot and wait for AI training to complete
6. Simulation: Run predictions based on the meeting conversation patterns

Natural language example:

"Search for all meetings about customer onboarding,
 create a mind from each unique customer contact"

AI automatically orchestrates all steps.

Topic-Based Mind Creation

Leverage Fireflies’ powerful search:

"Find all transcripts where pricing objections were discussed,
 create a mind trained on how our team handles pricing questions"

The AI will:

• Search transcripts for “pricing objections”
• Retrieve all matching transcripts
• Extract relevant segments
• Create a consolidated mind
• Train on pricing-related conversations

Tips for Effective Use

Search Strategies

By Keyword:

"Search transcripts for mentions of 'enterprise features'"
"Find meetings where we discussed Q4 targets"

Combined with Date:

"Search for 'customer feedback' in transcripts from last month"
"Find mentions of 'bug' or 'issue' in October meetings"

By Topic:

"Find all sales meetings"
"Search for product demo discussions"

Combining Tools

Get maximum value by using tools together:

"List all transcripts from this week, search them for
 mentions of our new feature, and if any customers gave
 feedback, create minds from those conversations"

This single command uses:

• fireflies_list_transcripts for discovery
• fireflies_search_transcripts for filtering
• fireflies_get_transcript for retrieval
• Mind Reasoner tools for mind creation

Unique Fireflies Features

Powerful Keyword Search

Unlike Gong and Fathom, Fireflies provides native keyword search:

• Search across all transcript text
• Find specific topics without knowing dates or participants
• Discover conversations you might have forgotten
• Build topical datasets for mind training

Speaker Analytics

Fireflies provides detailed speaker analytics:

• Talk time percentage
• Word count
• Speaking patterns
• Participation rates

Use these to:

• Identify dominant speakers
• Find balanced conversations
• Select high-quality training data

GraphQL Backend

Fireflies uses GraphQL (vs REST for Gong/Fathom):

• Efficient data fetching
• Flexible queries
• Rich data model

The MCP server handles all GraphQL complexity automatically.

Limitations & Considerations

Speaker Email Attribution

Critical limitation: Fireflies does not provide speaker emails directly.

How the MCP server handles this:

1. Fireflies provides speaker names (e.g., “John Smith”)
2. Fireflies provides meeting attendee list (names + emails)
3. MCP server correlates names to infer emails
4. Matching is best-effort, not guaranteed

Implications:

• Some speakers may not have email attribution
• Name variations can cause mismatches (“John” vs “John Smith”)
• External speakers may have less accurate matching

Solutions:

• Manually verify speaker emails when critical
• Use Gong or Fathom if accurate emails are required
• Review speaker attribution before training minds

API Rate Limits

Fireflies enforces API rate limits. The MCP server handles these automatically:

• Bulk operations may take time if rate limited
• Error messages will indicate when to retry
• Search operations count against rate limits

Data Access

You can only access:

• Transcripts you have permissions to view in Fireflies
• Transcripts from your organization’s Fireflies account
• Meetings that have been fully processed and transcribed

Next Steps