# Fireflies MCP for meeting intelligence. AI Agent Connect > Fireflies lets your AI client access your meeting history. Instead of manually digging through recordings, your agent can pull full transcripts, extract specific action items, analyze speaker sentiment, and use AskFred to answer questions about what happened in a call. It brings the intelligence from your Fireflies account directly to your preferred MCP-compatible client. ## Overview - **Category:** productivity - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_LWwgSkTY9PRaGPZKNZbB5DfsKtSW2mi1MV4q3M61/ai-agent-connect - **Tags:** fireflies, meeting-notes, transcription, action-items, askfred, meeting-intelligence, ai-notetaker, summaries, dialogue, team ## Description You don't need to rewatch hours of video to remember what was decided. This MCP gives your AI client a direct line to your Fireflies account. You can search through your entire meeting history using keywords or dates, then dive into the specifics of any call. Your agent can pull the exact list of action items, read the full dialogue line by line, or check the speaking metrics to see who dominated the conversation. If you need to verify a specific detail, your agent can find the exact timestamp and speaker for a quote. You can even use the AskFred functionality to ask questions about a single meeting or a whole series of calls. It turns your meeting recordings from static video files into a searchable, queryable database that your agent can use to keep your projects moving. ## Tools ### ask_fred Ask questions about a single meeting or a range of meetings. Your agent can poll for the answer if the response is still processing. ### get_meeting_action_items Extract the specific to-do list generated from a meeting. This includes the necessary context from the meeting outline. ### get_askfred_thread Retrieve the full history of a specific AskFred question and answer thread. ### get_meeting_dialogue Read the meeting transcript line by line. You can filter by speaker, search terms, or specific time windows. ### get_meeting_metrics Get data on speaking time, words per minute, and sentiment for everyone in the meeting. ### get_meeting Access a complete overview of a meeting, including attendees, the summary, and the topics discussed. ### list_askfred_threads View existing question threads from AskFred, which can be filtered by a specific meeting ID. ### list_meetings Search your meeting history by keyword, date, host, or participant. Results are returned newest first. ### list_team_members Get a list of all users within your Fireflies team account. ## Prompt Examples **Prompt:** ``` What was decided in the product meeting yesterday? ``` **Response:** ``` I found the product meeting from yesterday. The summary shows the final decision, and the action items list the specific owners assigned to each task. ``` **Prompt:** ``` Give me the action items from this morning's standup so I can put them in our tracker. ``` **Response:** ``` I pulled the action items from this morning's standup. They include the task descriptions and the relevant context from the meeting outline. ``` **Prompt:** ``` Find the exact moment the customer mentioned their budget and tell me what they said. ``` **Response:** ``` I found that moment in the dialogue. The customer mentioned the budget at the specific timestamp, and I have the exact quote for you. ``` ## Capabilities ### Transcript Retrieval Your agent reads the full text of any meeting to find specific details. ### Task Extraction The agent pulls action items directly from the meeting record. ### Speaker Analytics Your agent analyzes talk time and sentiment to evaluate meeting participation. ### Conversational Q&A The agent uses AskFred to answer complex questions about meeting content. ### Meeting Discovery Your agent searches through your history to find specific calls by date or participant. ## Use Cases ### Post-Meeting Follow-ups Your agent pulls action items immediately after a call to populate your task manager. ### Sales Debriefs An agent searches transcripts to find specific customer pain points or budget constraints. ### Meeting Audits Use speaking metrics to see if certain team members are being sidelined in discussions. ### Decision Tracking Search for specific keywords to find exactly when a project pivot was agreed upon. ## Benefits - Eliminates the need to manually rewatch meeting recordings. - Provides exact timestamps for quotes and decisions. - Converts spoken words into structured to-do lists. - Allows for cross-meeting analysis through AskFred. ## How It Works Connect your Fireflies account to the MCP and let your agent handle the heavy lifting. 1. Connect your Fireflies account using your own API key. 2. Your agent uses list_meetings to find the relevant conversation. 3. The agent pulls the transcript or action items using specific tools. 4. The data is delivered directly to your AI client for immediate use. ## Frequently Asked Questions **Do I need an API key?** Yes. Fireflies' GraphQL API authenticates every call. Generate a key at fireflies.ai → Settings → Integrations → API Keys, paste it into the credential field, and it is sent as "Authorization: Bearer ". **Can I read the actual spoken words, not just the summary?** Yes. get_meeting_dialogue returns the sentence stream with speaker, start and end time, and AI tags marking tasks, questions, prices, metrics, dates and sentiment. Filter it by speaker, a search string or a time window in seconds, then page through with offset/amount. **How do I find the meeting id Fireflies uses?** Run list_meetings with a date window, a host or participant email, or a keyword. It returns the transcript id alongside the title, date, duration and action-item count. Pass that id to get_meeting or any other tool. **What is the difference between get_meeting and ask_fred?** get_meeting returns the record Fireflies already built — the summary, action items, outline and attendees. ask_fred starts a new question thread in natural language, about one meeting (transcript_id) or across a range, and returns an answer plus suggested follow-ups you can continue with get_askfred_thread. **Can I see how much each person talked?** Yes. get_meeting_metrics returns per-speaker talk time and its percentage of the meeting, word count, words per minute, monologue count and longest monologue, filler-word count, plus the overall sentiment split and how many questions, dates, metrics and tasks were detected.