Granola (software)
Last reviewed
Sources
18 citations
Review status
Source-backed
Revision
v6 ยท 3,553 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
18 citations
Review status
Source-backed
Revision
v6 ยท 3,553 words
Add missing citations, update stale details, or suggest a clearer explanation.
Granola is an artificial intelligence-powered meeting notepad application that combines a user's own typed notes with AI-generated content derived from a meeting's audio transcript. Unlike conventional AI meeting assistants that deploy a visible bot into video calls, Granola captures audio directly from the user's device, making it invisible to other meeting participants. The software was founded in March 2023 by Chris Pedregal and Sam Stephenson and is headquartered in London, United Kingdom [1][2].
Granola's core approach positions the user as an active participant in the note-taking process rather than a passive recipient of AI-generated summaries. During a meeting, the user jots down brief notes about what they consider most important, and Granola's AI then enriches those notes with details from the full transcript once the meeting ends. The company describes this as providing a "steering wheel for LLMs," allowing users to guide the AI's output rather than relying on fully automated summarization [3].
As of February 2026, Granola has raised $67 million in total funding across three rounds and is valued at $250 million. The company employs approximately 87 people [2][4].
Granola was co-founded by Chris Pedregal and Sam Stephenson in March 2023. Pedregal, who studied computer science at Stanford University and holds an MA in journalism from the same institution, is a serial entrepreneur with a track record of building companies acquired by Google. His first company, Apture, a contextual browsing startup he co-founded while at Stanford, was acquired by Google in November 2011 to enhance the Chrome browser [5]. He then worked as a product manager at Google, contributing to Gmail, Google Search, and Google Maps.
In 2013, Pedregal co-founded Socratic with Shreyans Bhansali. Socratic was an AI-powered educational application that allowed students to photograph homework problems and receive step-by-step explanations. Google acquired Socratic in March 2018 and relaunched the app on iOS in August 2019 under the Google brand [6]. Pedregal was named to Forbes 30 Under 30 in Education for his work with Socratic.
After the Socratic acquisition, Pedregal served as Product Lead for Stack at Google from April 2020 to June 2022. Stack was a document scanning and organizing tool that was integrated into Google Drive. After leaving Google in 2022, Pedregal moved to London with his family and began working on what would become Granola [2].
Sam Stephenson, Granola's co-founder and head of design, studied graphic design at Falmouth University before spending several years in San Francisco working at a design agency and an education nonprofit. He had previously built a startup connecting neighborhoods to local farmers and worked on design projects for companies including Carv (interactive ski maps) and Swim Smooth (an iOS app). Stephenson and Pedregal spent months prototyping various work tools with test users before settling on the meeting notes concept that became Granola [7].
Granola closed a $4.25 million seed funding round in 2023, led by Lightspeed Venture Partners. Other investors in the seed round included Betaworks, Firstminute Capital, Otherwise Fund, Uncommon Creative Studio, and various angel investors [3].
The company publicly launched its product in May 2024 with a blog post titled "Introducing Granola: The AI Notepad That Enhances, Not Replaces, Your Thinking In Meetings." The launch emphasized Granola's philosophical departure from existing AI meeting tools. Rather than attempting to replace human note-taking entirely, Granola was designed to augment it, using the user's own notes as a signal for what matters most and then filling in details from the transcript [1][3].
At launch, the product was available only on macOS. The application appeared as a simple notepad window that users could open alongside their video conferencing tool. During a meeting, Granola would capture audio from the system's microphone and speakers while the user typed informal notes. When the meeting ended, the AI combined these inputs to produce structured, detailed notes [3].
In October 2024, Granola raised a $20 million Series A round led by Spark Capital, with participation from AI Grant, Lightspeed Venture Partners, Betaworks, and Firstminute Capital. At the time of the raise, Granola had approximately 5,000 weekly active users. The company had maintained a 10% week-over-week growth rate in active users since its May 2024 launch, with 50% user retention at ten weeks [8].
The Series A funding was used to expand the engineering team, improve the AI models underlying note generation, and begin development of team collaboration features. The round brought Granola's total funding to approximately $24 million [8].
In May 2025, Granola raised a $43 million Series B round led by NFDG, the venture firm of Nat Friedman (former CEO of GitHub) and Daniel Gross. Existing investors Lightspeed Venture Partners and Spark Capital also participated. Angel investors in the round included Guillermo Rauch (CEO of Vercel), Amjad Masad (CEO of Replit), Tobias Lutke (CEO of Shopify), and Karri Saarinen (CEO of Linear). The round valued Granola at $250 million and brought total funding to $67 million [4][9].
Alongside the Series B announcement, Granola launched Granola 2.0, which transformed the product from a personal AI notepad into a collaborative workspace for teams. The update introduced shared team folders, cross-meeting search powered by reasoning models, Slack integration for automatic meeting summaries, and enterprise-level collaboration features [4].
At the time of the Series B, the company had 19 employees and planned to grow to approximately 80 within 12 months. By February 2026, the headcount had reached 87 [2][9].
Granola's most distinctive technical characteristic is its bot-free architecture. While competitors like Fireflies.ai, Otter.ai, and Fathom typically deploy a visible bot participant into video calls (appearing as "Fireflies.ai Notetaker" or similar), Granola runs entirely on the user's local device. It captures audio through the system's microphone and audio output without joining the meeting as a separate participant [10].
This approach has several practical implications:
| Aspect | Bot-based approach (competitors) | Granola's bot-free approach |
|---|---|---|
| Visibility | Bot appears in participant list; all attendees can see it | Invisible to other participants |
| Meeting dynamics | Participants may change behavior when they know a bot is recording | Participants behave naturally |
| Platform compatibility | Often limited to specific platforms (Zoom, Teams, Meet) | Works with any meeting software that produces audio |
| Admin permissions | May require IT admin approval to allow bots into calls | No admin permissions needed |
| Recording indicator | Clear recording notification via bot presence | No external indicator of recording |
Granola works across Zoom, Google Meet, Microsoft Teams, WebEx, Slack Huddles, and any other software that produces audio on the user's device. It also supports in-person meetings (by capturing microphone audio) and can transcribe voice memos, podcasts, or YouTube videos [10].
Granola's workflow follows a three-stage process:
Before the meeting: Granola detects an upcoming meeting from the user's calendar and presents a blank notepad window. Users can optionally select a meeting template to structure the output.
During the meeting: The user types brief, informal notes about what they find most important. These can be as minimal as a few bullet points or keywords. Meanwhile, Granola transcribes the meeting audio in the background. The user's notes appear in black text, while AI-generated content is shown in gray, maintaining a clear visual distinction between human and machine contributions.
After the meeting: When the call ends, Granola combines the user's notes with the full transcript and uses large language models to produce structured, detailed meeting notes. The AI prioritizes the topics the user flagged as important in their own notes, then fills in supporting details from the transcript. If the user writes nothing during the meeting, Granola still generates notes based solely on the transcript [1][3].
This hybrid approach means the resulting notes reflect what the user thought was important, backed by the full record of what was actually said. Granola's team has described this as giving users a "steering wheel" for the AI, contrasting it with fully automated tools that force users to accept whatever the algorithm decides is relevant [3].
Granola uses speech recognition to transcribe meeting audio in real time. The transcription runs locally on the user's device during the meeting, then is processed by cloud-based AI models after the meeting ends to generate the enhanced notes. Audio is temporarily cached during the meeting for transcription purposes; once transcription is complete, the audio is deleted from Granola's systems and any third-party services. Audio recordings are not retained [11].
The platform supports transcription in multiple languages, with multi-language support available across all pricing tiers [12].
Granola offers pre-configured templates for common meeting types and allows users to create custom templates. Templates define the structure and content categories that the AI uses when generating notes.
| Template type | Extracted content |
|---|---|
| Sales call | Pain points, feature requests, objections, next steps, deal stage signals |
| Customer discovery | Customer problems, use cases, feature feedback, follow-up actions |
| Job interview | Candidate responses, strengths, concerns, hiring recommendation |
| 1:1 meeting | Discussion topics, decisions made, commitments from both parties |
| Sprint planning | Sprint goals, task assignments, estimated effort, blockers |
| Product review | Feature assessments, user feedback themes, prioritization decisions |
| Weekly team meeting | Updates by team member, blockers, action items, announcements |
| Investor update | Key metrics, milestones, challenges, asks |
Users can create their own custom templates to match specific meeting formats, and these templates can be shared across an organization to standardize note-taking. Granola automatically suggests the most relevant template for each meeting based on the meeting title, description, attendee roles and departments, and patterns from previous meetings [13].
Selecting a different template after a meeting has ended regenerates the notes using the new template's structure, allowing users to reformat notes without losing any information [13].
Granola Chat allows users to ask natural language questions about their meetings. The feature works at two levels: users can query a single meeting's transcript for specific details, or search across all meetings in a folder or workspace for broader patterns and information [14].
Typical queries include:
Responses include citations that link back to the specific meetings where the information was found, allowing users to verify the AI's answers against the original transcript. The scope can be set to query notes only or to use full transcripts (limited to the 25 most recent meetings). Chat messages are not saved between sessions [14].
Introduced with Granola 2.0 in May 2025, shared folders allow teams to organize meeting notes into dedicated spaces such as "Sales Calls," "Customer Feedback," "Hiring Loops," or "Weekly Syncs." Team members can access folders even without a Granola account. The folder-level chat feature allows users to query across all notes within a folder, with the AI citing specific meetings as sources in its responses [4].
Business and Enterprise users can browse any public folder within their organization's domain, which is useful for competitive intelligence, customer success tracking, or onboarding new team members who need to catch up on past discussions [4].
Granola integrates with a range of business tools to route meeting notes and action items into existing workflows.
| Category | Integrations |
|---|---|
| CRM | HubSpot (native), Attio (native), Affinity (native), Salesforce (via Zapier) |
| Collaboration | Slack (native), Notion (native) |
| Automation | Zapier (connects to 8,000+ apps via two triggers: "Note Added to Folder" and "Note Shared to Zapier") |
| AI assistants | Model Context Protocol (MCP) for Claude, ChatGPT, and other MCP-compatible tools |
| Project management | Linear, Asana, Jira, Monday.com (via Zapier) |
The HubSpot integration (available on Business and Enterprise plans) syncs enhanced meeting notes directly to HubSpot Contact records with a single click. The sync includes the meeting title, date, participant list, the enhanced note body, and any extracted action items. Custom template sections are preserved as formatted text in the HubSpot activity note [15].
For Salesforce users, the Zapier integration provides an indirect connection, though there is no native Salesforce integration [15].
Granola offers a Model Context Protocol (MCP) server that allows AI tools such as Claude, ChatGPT, and code editors like Cursor to access meeting data directly. Through MCP, developers can ask Claude to create tickets for bugs discussed in a meeting, scaffold features based on technical discussions, update a Linear board from a standup, or draft and share sales call notes to a CRM [16].
The MCP server operates on local cache files, reading meeting data directly from Granola's local cache without making API calls to Granola's servers. This design provides offline capability, fast access without API rate limits, and a privacy-focused architecture where meeting data does not leave the user's machine [16].
Granola is available on macOS, Windows, and iOS.
| Platform | Requirements | Capabilities |
|---|---|---|
| macOS | macOS 13 minimum; macOS 14.2+ recommended | Full feature set: virtual meetings, in-person meetings, voice memos, any device audio |
| Windows | Windows 10+ | Full feature set: virtual meetings, in-person meetings, voice memos, any device audio |
| iOS (iPhone) | iOS app | Designed for in-person meetings; captures audio via device microphone |
There is no Android app or web-based version as of early 2026. On desktop (macOS and Windows), Granola can transcribe and summarize meetings on any video conferencing platform, in-person meetings, voice memos, and any audio playing on the computer such as podcasts or YouTube videos [10].
Granola uses a freemium pricing model with three tiers.
| Plan | Price (per user/month) | Key features |
|---|---|---|
| Basic (Free) | $0 | AI-enhanced note-taking, Granola Chat, shared folders, multi-language support, 30-day note history |
| Business | $14 | Unlimited meeting notes and history, advanced AI models, advanced integrations (HubSpot, Attio, Slack), centralized billing, user management |
| Enterprise | $35+ | Single Sign-On (Okta, Google Workspace), organization-wide AI training opt-out, usage analytics, priority support, API access, configurable data retention |
The free Basic plan provides access to core features but limits users to viewing only the last 30 days of notes. The Business plan removes this restriction and unlocks advanced integrations and AI models. The Enterprise plan adds compliance and governance features required by larger organizations [12].
Granola's security architecture reflects its bot-free design philosophy. All data is stored on Amazon Web Services (AWS) servers located in the United States, with encryption applied both at rest and in transit. Audio is temporarily cached during the meeting for transcription; once transcription is complete, the audio is deleted and is not retained by Granola or any third-party services [11].
Notes and transcripts are retained indefinitely unless configured otherwise. Enterprise plans offer configurable auto-deletion retention periods for transcripts [11].
| Certification/Standard | Status |
|---|---|
| SOC 2 Type II | Achieved July 2025 |
| GDPR | Compliant; Data Processing Agreement available |
| UK GDPR | Compliant |
| Data residency | US only (no EU/UK regional data residency option) |
Granola's GDPR compliance includes a dedicated Data Protection Officer, clear lawful basis for processing, data minimization practices, right to erasure, data portability under GDPR Article 20, and privacy-by-design principles [11].
Because Granola does not announce its presence in meetings (no bot joins the call), the responsibility for informing meeting participants about recording and transcription falls entirely on the user. In two-party consent or all-party consent jurisdictions (such as California, Illinois, and parts of the European Union), users are legally required to inform all participants and obtain consent before starting a Granola session. The tool's invisibility does not change the legal baseline for recording consent [17].
Granola competes in the AI meeting assistant market, which has grown significantly since the shift to remote work during the COVID-19 pandemic. The company differentiates itself from competitors primarily through its bot-free recording approach, its hybrid human-AI note-taking model, and its design-forward user interface.
| Competitor | Approach | Key differentiator |
|---|---|---|
| Otter.ai | Bot joins meeting; real-time transcription | Live collaboration on transcripts; conversational search |
| Fireflies.ai | Bot joins meeting ("Fred") | Cross-platform support (69+ languages); AskFred conversational AI; sentiment analysis |
| Fathom | Bot joins meeting | Generous free plan; fast AI summaries; popular with individual users |
| tl;dv | Bot joins meeting | Multilingual transcription (30+ languages); snippet sharing for async review |
| Read.ai | Bot joins meeting | Meeting engagement scoring; real-time coaching alerts |
| Gong | Bot joins meeting | Revenue intelligence and deal forecasting for enterprise sales teams |
| Fellow | Bot joins meeting | Meeting management with agenda setting, action item tracking, and one-on-one templates |
Granola's primary competitive advantage is its invisible operation. By not deploying a bot, Granola avoids the social friction that occurs when participants notice a recording bot has joined their call. This is particularly valued by venture capitalists, consultants, and professionals in sensitive client-facing meetings where a visible recording bot could alter the conversation's tone or make participants uncomfortable [17].
However, the bot-free approach also has trade-offs. Bot-based competitors can capture video along with audio, provide real-time transcription visible to all participants, and offer recording capabilities that do not require a desktop application. Granola's local audio capture requires the application to be installed on the user's computer, whereas bot-based tools can work from any device since the bot joins independently [17].
The hybrid note-taking model (combining user notes with AI transcript analysis) is another differentiator. Most competitors generate notes entirely from the transcript without user input, which can result in summaries that miss what the user actually cared about. Granola's approach produces notes that are personalized to the user's priorities while still capturing the full breadth of the conversation [3].
Granola has been well-received among product managers, venture capitalists, sales professionals, and consultants. The product's design, which mimics a simple notepad rather than a complex meeting management platform, has been praised for reducing the cognitive overhead of using AI tools during meetings. Users have noted that the ability to type informal notes without worrying about structure or completeness, and then receive polished output after the meeting, reduces the anxiety associated with note-taking in fast-paced conversations [18].
The product has attracted a user base that skews toward professionals who are in back-to-back meetings throughout the day, where the ability to quickly capture key points and let AI handle the details is particularly valuable. The company maintained a 10% week-over-week growth rate in active users following its May 2024 launch and reported 50% user retention at ten weeks [8].
Critics have pointed to several limitations. The lack of video recording (since there is no bot to capture video) means users who need visual records of meetings must look elsewhere. The requirement for a desktop application (no web version) limits flexibility, and the absence of an Android app excludes a segment of mobile users. Some reviewers have also noted that while Granola excels for individual note-taking, its team collaboration features (introduced with Granola 2.0) are still maturing compared to more established competitors [17][18].