Skip to content
MakerGain
← Back to project

In the Years Fathom Made Meeting Notes Free, the Founder Says It Lost \$50 per User per Month

Richard White built Fathom to keep the thoughts from his meetings: the personal tier offers unlimited recording and transcription for $0, while the Team tier charges for sharing and search and the Business tier charges for CRM write-back. A HubSpot example on the official site requires proposed field changes to be shown first and moves a stage only based on the buyer’s own words; White said Fathom lost about $50 per user per month in its early years, and in September 2026 Fathom became part of Superhuman.

Richard White found that after meetings, his ideas and decisions rarely made it into notes or follow-ups. He built himself a tool that recorded conversations, generated notes, and kept the action items. That origin is what Fathom’s About page describes. (About[1])

The Fathom that followed is an AI meeting notes tool that generates a summary as soon as a call ends, syncing notes, insights, and action items to Slack, Salesforce, HubSpot, Notion, and Asana. The personal version is $0, free forever, with unlimited recording and transcription. (pricing page[2])

A stranger arriving here will ask: if it is unlimited and free, where does the money come from? In an interview in May 2026, White gave a number—in Fathom’s first few years, it lost about $50 per user per month. That is his account of a historical stage, not today’s unit cost, and it does not imply profit. (interview[3])

“Only Move the Stage If the Buyer Actually Says So”

Start with the work that happens after a meeting. After a sales call, updating the deal record in HubSpot is the first thing to get delayed. Fathom’s website has a HubSpot work example that breaks this down in detail: after the call ends, ChatGPT or Claude reads the conversation through Fathom, finds the matching deal in HubSpot, proposes updates to the deal stage, next steps, pain points, and tasks for both sides, and then drafts a follow-up email for review. (HubSpot example[4])

The rules are more interesting than the features. Before writing to HubSpot, every proposed field change must be shown, and pain points must be supported by the customer’s own words; before creating a new deal or contact, the system asks first; most important—only move the deal stage if the buyer actually said words that support the stage change.

That rule leaves the decision with people: “I think we can move this forward” and “the customer agrees to move forward” are two different voices in a meeting, and who gives the nod is decided by someone who understands the context. It does not sell “fully automated AI” as the draw; instead it adds a confirmation step to the process.

This is the vendor’s own example, with no independent customer results and no validation of field-write accuracy. The FAQ on the same page says any plan, including Free, can read calls in ChatGPT or Claude via MCP; Fathom’s existing HubSpot integration syncs call summaries over, and this example adds field updates and a follow-up draft on top. The Business tier on the pricing page has a separate native “CRM field sync” that automatically updates records after meetings—a different path from this MCP example. (pricing page[2])

Image: Fathom official showcase[5], accessed 2026-11-10; original interface preserved.

The Free Tier Makes Recording the Default; the Team Tier Carries the Paid Case

Fathom’s pricing turns “recording” into the default action: the personal Free tier is $0, free forever, with unlimited recording and transcription, instant AI summaries, clips and playlists, and cross-call search. Above the personal tier is the paid Premium tier, which adds advanced summaries, AI action items, and a custom meeting bot. (pricing page[2])

The Team tier has a clearer boundary. Team is priced per seat, with a minimum of 2 people, and lets a team put call records in shared folders, search and ask questions across multiple calls, and includes team security controls and HIPAA compliance under a standard BAA. The pricing page lists Team as $19 and $15, with monthly and annual options, but the visible text does not fully label which number corresponds to which. The Business tier adds CRM field auto-sync, Deal View, and coaching scorecards. For a sales team, the reason to pay is not “it can record” but “the records can be shared across the team, queried, and written back to the CRM.”

In the same interview, White said the team did things in order of risk: first onboarding and activation, then acquisition, then referrals, and only last monetization. He also recalled that the team made a pitch deck first, saying they were “selling what amounted to a roadmap.” (interview[3])

There are also two moves that do not look like standard SaaS: they set a principle of reserving 15% of every funding round for user investment; on distribution they opened several channels—eligible startups affiliated with VCs or accelerators get up to two years of free Team, nonprofits get 10 free seats, and those switching from products such as Gong get Business free for the term of their contract, along with data migration. (pricing page[2]) The channels are laid out here; public materials do not say which one actually brought in how much paid business.

There is one revenue number to be careful with. The interview page title says Fathom reached $30M ARR, while the body quotes White saying that in the first three years of monetization it went from zero to $30M, and a field on the same page labels it as Company revenue. This is an interview self-account and a page label, not audited financials; paid seat counts, ARPU, churn, and free-to-paid conversion are all undisclosed. (interview[3])

The Record Itself Is Hard to Stand Alone as an Endpoint

On September 14, 2026, White announced that Fathom had officially become part of Superhuman. He said that on that day users’ apps, logins, and product experience were unchanged, and that the Fathom connector in Superhuman Go was already live, able to query summaries, transcripts, action items, and recordings from past meetings. (announcement[6]) The announcement did not include the deal price, structure, or post-integration product roadmap.

This ownership somewhat echoes the product’s original form: meeting notes themselves are hard to stand alone as an endpoint; they are more like a piece of source material that others read when making judgments. In the announcement, White wrote that the next step was “less about searching and more about acting”—letting the right context find the people who need it, rather than people digging through call records. The homepage also says meeting data can be used in external assistants such as ChatGPT and Claude. (Fathom website[5])

There is another approach on the market. Granola explicitly does not use a meeting bot, relying on computer system audio and the microphone, and can handle scheduled, ad hoc, online and offline conversations; it processes audio for transcription but does not record or retain audio. Not using a bot means it is not automatically visible to other participants, so it provides automatic meeting chat messages and video watermarks to let participants know Granola is transcribing; reminding users to obtain consent is another mechanism—each time they start taking notes, Granola gives a reminder. On price, the Basic tier is $0, with meeting history retained for only 30 days; the Business tier is $14 per user per month, with unlimited notes and history and includes MCP and API. (Granola explanation[7], pricing[8])

These two are not mutually exclusive categories; Fathom itself also has bot-free capture, though the pricing page labels it as a beta feature on Mac. “Whether audio is retained” is a real choice; Fathom’s public materials do not lay out complete audio and transcript retention periods or deletion policies, so the comparison can only go this far. (pricing page[2])

White initially just wanted to keep the thoughts that would disappear from his meetings. Fathom turned recording into a free default action, and made team sharing and CRM write-back the reasons to pay; whether a human has to give the nod when writing back, the MCP example clearly says yes: every field change is shown first, then saved. Business’s native sync only says it automatically updates records; public materials do not say whether it also stops for confirmation. That calibration is closer to the real office than the “fully automated AI” slogan.

Continue reading

AI search
He Treated Reddit as a Survey. 135,000 People Used It for Research. In 2025, It Stopped at a License

Fed built GummySearch into a tool for audience research using public Reddit discussions: it searched and tracked community conversations by categories such as pain points and people seeking solutions. After commercial licensing talks with Reddit failed, he chose to stop sales on November 30, 2025, and gradually shut down the site rather than continue operating without a license; paying users kept transitional access, and the free tier retained only login, export, and account deletion.

AI productivity
The day the accountant came asking for the bookkeeping records, he decided to let AI go through his own inbox

In March 2023, software company founder Romeo Bellon was overdue on bookkeeping and his accountant pressed him for records. He developed Receiptor AI, which scans inboxes, extracts receipts and invoices, and pushes them to accounting software; he launched it with Luigi Fernandez Ortega in August. It turns historical email into a retroactive scan priced by email volume, while subscriptions process only new email; its privacy policy says documents are retained indefinitely and automatic deletion is still in development.

AI audio
The Source Code Is Out There; the Installed Version Sells for \$25: VoiceInk Puts Dictation in Any Text Field

VoiceInk is an instant-dictation app for Apple Silicon Macs: press a shortcut to write speech into a text field, or display an AI answer or run a command. Its source code is public under GPLv3; the prebuilt version is a one-time purchase at $25, $39, or $49, licensed by device count and including updates and support; cloud services require your own API key. Its affiliate program pays 30% commission but bans paid traffic, and the pricing page and README disagree on the macOS requirement.

AI audio
MacWhisper Turns Recordings Into Text, Then Leaves Identifying the Speakers to You

Podfeet's Allison uses MacWhisper for interview captions and transcripts: she enters a 12.6-second offset, exports an SRT, then labels speakers sentence by sentence and merges paragraphs. Automatic recognition can separate voices, but it cannot decide identities for her; she uses ChatGPT to help with segmentation and runs into confusion between an API key and a Plus subscription. Developer Jordi Bruin tightens the terms because of license resale.