Fireflies + AI: Use Fireflies in Juma's AI Marketing Workspace
Connect Fireflies to Juma to turn meetings into client recaps, voice-of-customer libraries from Soundbites, and account briefings through chat, plus light write actions like cutting soundbites and managing meeting access, in the team's brand voice.
Key takeaways
Yes, Fireflies works with AI through Juma
Connection runs on Fireflies's official MCP (Model Context Protocol) server, the first AI-meeting-tool MCP shipped to a public AI assistant.
Read plus light write through chat
Draft post-call client recaps and follow-ups, build voice-of-customer libraries from Soundbite clips and transcripts, prepare account briefings for upcoming meetings, and take write actions like cutting a soundbite from a transcript, sharing or renaming a meeting, or moving it to a channel.
Project-aware drafting
Meeting context blends with the team's brand voice, persistent Project knowledge, and other connected tools so recaps and briefs sound like the team wrote them, not a generic AI.
Works alongside HubSpot, Slack, Notion, and the rest of the marketing toolchain
In one chat. Fireflies on its own answers about meetings; Juma blends meetings with the rest of the team's context.
Used by 400+ marketing teams
Including Salesforce, Maersk, Costa Coffee, Wavemaker, and Mindshare. 4.9/5 on G2.
Enterprise-grade by default
SOC 2 Type II, ISO 27001, GDPR, HIPAA. AES-256 encryption. Zero data retention from AI providers.
Can I access Fireflies through AI?
The connection runs on Fireflies's Model Context Protocol (MCP) server. MCP is an open standard that lets AI assistants call external tools; Fireflies shipped first-party MCP in June 2025 and was the first AI meeting tool listed in Claude's official MCP Directory in August 2025, several months ahead of competing meeting platforms. Juma uses the same MCP rail, so the underlying API access is identical to connecting Fireflies to Claude or ChatGPT directly.
The connection runs at the user level: each user authorizes Juma with their own Fireflies credentials, so the meetings Juma can see match that user's existing Fireflies permissions and channel access. Once connected, Fireflies is available across every Juma project and chat, with no per-project remapping. Fireflies is most adopted by sales, CS, and enterprise teams (notably for its conversation analytics, 200+ scoped AI Apps, and SOC 2 + HIPAA + GDPR compliance depth), but the jobs apply equally to agencies and in-house marketing for client recaps, voice-of-customer libraries, customer-research interviews, and the sales-to-marketing handoff.
What can Juma do with Fireflies?
Through Juma the team can draft post-call recaps and follow-ups, build voice-of-customer libraries from Soundbite clips, prepare account briefings for the next meeting, and take light write actions like cutting a soundbite or managing meeting access, all from chat. Outputs land as ready-to-send emails, structured synthesis docs, sharable account briefs, or a new soundbite clip, in the team's brand voice from Project knowledge.
01 · Post-call client recap
Post-call client recap and follow-up drafting
Ask in chat: "Draft a follow-up email for yesterday's Acme call in our voice, with action items, owners, and a next-meeting suggestion." The summary, decisions, and action items come from Fireflies, client context from the matching Project, and the recap comes back ready to send. Strategy, taste, and judgment stay human; the writing scaffolding is handled. The same chat drafts the QBR deck when the question shifts from "what did we discuss today" to "what did we discuss last quarter."
02 · Voice-of-customer synthesis
Voice-of-customer synthesis and Soundbite library curation
Ask in chat: "Find every Soundbite tagged 'pricing objection' from the last 30 days, summarize the recurring objection language, and suggest three positioning angles for the next campaign." The Soundbite library and underlying transcripts come back as a synthesis with quoted evidence plus positioning recommendations. PMM teams use this for enablement collateral in customers' own words; content teams pull verbatim phrasing for landing-page copy; account leads spot recurring asks before they become churn risks.
Fireflies Soundbites are audio clips, not video, so the synthesis surface is text plus audio attribution. Teams that need on-camera testimonial video should plan for a separate capture path.
03 · Account briefing
Account briefing for the next meeting
Ask in chat: "Brief me on every call with the Acme team this quarter: open action items, decisions, sentiment trend, and what to lead with next time." Prior meetings come back as a structured briefing with open threads, recommended openers, and a sentiment trend from Fireflies's conversation analytics. Agency account managers, CS leads, and sales reps walk into the next call already aligned. For agencies running many clients in parallel, the briefing adapts to the Project so each client's brief reads like the account lead wrote it.
04 · Create soundbites
Create soundbites and manage meeting access
Ask in chat: "Cut a soundbite from yesterday's Acme call between the 12:30 and 13:10 marks where they describe the pricing objection." Given the moment, Juma creates the soundbite clip from the transcript. The team can also share a meeting with specific emails, revoke that access, set its privacy, rename it, or move it to a channel or folder, all from chat. These are the write actions Fireflies exposes through its MCP server: Juma never reaches past them into a CRM or another system of record. Strategy and judgment stay human; the team confirms each share or privacy change before it takes effect.
How does Juma + Fireflies replace your meeting-notes stack?
Most teams stack Fireflies plus a separate recap-writing AI, a slide tool for QBRs, manual CRM data entry, a spreadsheet for cross-call themes, and a prep-doc tool. With Juma + Fireflies, the team keeps Fireflies as the capture and analytics layer; drafting, synthesis, scoring, and prep happen in one chat, in the team's brand voice from Project knowledge.
Juma augments the team, it doesn't replace it. What disappears is the layer between Fireflies's transcripts, summaries, Soundbites, and conversation analytics and a finished deliverable: the writer turning summaries into recaps, the analyst building the QBR deck, the spreadsheet for cross-call themes, the manual clipping of soundbites by hand. Customers report 60% faster workflows and 50+ hours saved per month once Juma sits between source tools and finished output, with the team reviewing every draft before it ships.
What does Fireflies work best with in Juma?
Three pairings make the most of Fireflies data through Juma: pair Fireflies with HubSpot to read both the call and the deal context and draft a grounded follow-up, pair it with Slack to surface voice-of-customer moments to the team that needs to hear them, and pair it with Notion to build a persistent knowledge artifact from per-call data.
Read the call and the deal, then draft the follow-up
Fireflies answers what did the prospect or client actually say, and how (talk ratio, sentiment, objection language, Soundbite clips). HubSpot answers what's the current state of this contact, deal, or account in our system of record? Both connections are read-only on the CRM side, so Juma reads the call from Fireflies and reads the deal and contact context from HubSpot, then drafts a follow-up email or call recap that reflects both. The draft comes back ready for the team to review and send; Juma does not write back to HubSpot. Agencies running their own and their clients' funnels lean on this read-and-draft loop to keep recaps grounded in the live deal state.
Voice-of-customer distribution to the team
Fireflies answers what was said and which 30-second Soundbite captures it. Slack answers where does the team see new context as it happens? Pair them to route Soundbites, summaries, and scoped AI-extract outputs (BANT scorecards, objection-handler briefs) into the channels where copywriters, strategists, and account leads already live. PMM teams pull live objection language for enablement collateral; agency creative leads spot client-call moments without logging into Fireflies.
Persistent knowledge artifact from per-call data
Fireflies answers what was said in this one call (or across this client's calls this quarter). Notion answers what do we collectively know about this client or topic, accumulated over time? Pair them to maintain a per-account or per-research-program Notion page that accumulates Fireflies summaries, Soundbite embeds, AI App output, and AskFred answers. Agencies use this for client account docs; PMM and content teams compile interview synthesis into a durable research base.
How to connect Fireflies to Juma
Open Integrations
Click your name in the bottom-left corner of Juma, then Integrations.
Find Fireflies
Under Meetings & Communication integrations.
Connect
Click Connect and sign in with your Fireflies credentials. The connection runs at the user level: the meetings, Soundbites, and channels the team member can see in Fireflies are what Juma can see through them.
Start using Fireflies
Available across every Juma project and chat. Ask directly in chat: "Draft a follow-up email from yesterday's Acme call" or "Find Soundbites tagged 'pricing objection' from the last 30 days and summarize the patterns."
Trusted by marketing teams that ship
Customer outcomes
240+ teams
Security & compliance
Questions, answered
Can I use Fireflies with ChatGPT or Claude instead of Juma?
Yes. Fireflies shipped a first-party MCP server in 2025 and is supported in Claude, ChatGPT, Cursor, GitHub Copilot, and Microsoft Copilot, so any user can pull meeting context into those tools directly. What Juma adds is the workspace around the meetings: one Project per client with persistent brand knowledge, multi-tool composition (Fireflies alongside HubSpot, Slack, Notion, GA4), finished deliverables instead of chat responses, and platform-level controls (audit logs, SSO/SAML, SOC 2 Type II). For a single user, raw Claude or ChatGPT works; for a team running many client accounts, the Project layer matters.
How much does it cost to use Fireflies through Juma?
Juma prices by what gets done, not by which model ran. Tasks consume credits: 1 to 100 for low-effort tasks, 100 to 500 for high-effort tasks like QBR drafting or quarterly VOC syntheses, and zero for failed tasks. Free, Pro, and Enterprise plans all include unlimited seats and access to Fireflies through the connector; they differ only in monthly credit volume. The Fireflies subscription is separate; Juma does not resell Fireflies and adds no markup on Fireflies's data.
What's the difference between AskFred and Juma's Fireflies integration?
AskFred is Fireflies's native query layer: ask questions about meetings inside the Fireflies app, with answers cited to timestamps in the underlying transcript. Juma's Fireflies integration sits one layer up: meeting context blends with the team's brand voice, Project knowledge, and other connected tools (HubSpot, Slack, Notion) inside a marketing workspace. AskFred answers about the calls; Juma turns the answer into a finished asset (client recap, VOC synthesis, account brief, BANT scorecard) in the team's voice.
Does the Fireflies + Juma integration support agencies managing multiple clients?
Yes. Projects organize knowledge per client (brand voice, prior decisions, asset library), and the Fireflies connection works across every Project the user touches, subject to Fireflies channel/folder permissions. 400+ marketing teams use Juma, including agencies like Wavemaker, Mindshare, and Mission North. SOC 2 Type II and ISO 27001 compliance covers client data handling; Enterprise adds SSO/SAML and private cloud. Unlimited seats on every plan means the whole agency works on the same Fireflies data without per-user fees.
Can we cancel Fireflies and just use Juma?
No. Juma needs Fireflies to capture, transcribe, and structure the calls (including Soundbites, conversation analytics, and the 200+ scoped AI Apps Fireflies runs inside its own product). Fireflies stays as the meeting layer. What Juma replaces is the stack between Fireflies's summaries, transcripts, Soundbites, and analytics and a finished deliverable: the separate recap-writing AI, the QBR slide builder, the spreadsheet for cross-call themes, and most of the analyst hours spent stitching the pieces together.
Does Juma store my Fireflies data?
Fireflies data is pulled on demand: only the meetings, transcripts, Soundbites, and analytics a specific question needs come through, not bulk export. Results stay inside the Juma workspace, encrypted with AES-256 in-transit and at-rest. Juma's zero data retention policy with AI model providers means raw Fireflies data is never used to train any third-party model. SOC 2 Type II and ISO 27001 audits are available on request. Fireflies can be disconnected at any time without affecting the rest of the workspace.
Can Juma create video clips from Fireflies recordings?
No, with a caveat. Fireflies Soundbites are audio-only by design, so the clip surface Juma works with is audio plus the underlying timestamped transcript. Teams that need on-camera testimonial video should plan for a separate capture path. Voice-of-customer text, quote banks, and audio clips for podcast or social-audio use are where Fireflies's curation model leads, and Juma builds on top of that.
Can Juma run Fireflies's 200+ scoped AI Apps from chat?
Not directly. Fireflies's AI Apps (BANT, Objection Handler, Brand Voice Checker, and the other 200+) run inside Fireflies and produce structured outputs there. Those outputs are pulled through the integration and folded into recaps, briefs, and synthesis docs in Juma. From a Juma chat the team can also run scoped prompts against Fireflies transcripts and summaries that mirror the shape of the AI App outputs, using Project knowledge to ground the result in the team's brand voice.
Ask AI about us
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
What's on your plate today?
Start typing a task and Juma will figure out the rest.