Tome

» see tome.app as it lived, on the wayback machine

~ the obituary ~

tome was the kind of thing the y-combinator timeline got excited about for a stretch. that was not my timeline. honestly, by 2022 when the ai pivot landed i was a few years into the edge-vision work in ankara and the part of twitter where founders posted shipped-this-week threads about ai slide tools was somewhere i did not really hang out. anyway, the only time tome showed up in my world was as a side panel in a tech-news email i half-skimmed on the morning train.

keith peiris and henri liriani, both ex-instagram and ex-facebook, opened tome in september 2020 as a story-format slide builder, not yet an ai product. the original pitch was that powerpoint and google slides were locked into the 16:9 page metaphor while the actual surfaces where slides got consumed had moved to mobile and to the web (more or less). tome built a card-stack format that scrolled vertically, embedded interactive blocks. in those days. live data from notion and figma and looker.

the ai pivot landed in september 2022, roughly the same week chatgpt entered the wider zeitgeist. tome added a generate-deck-from-a-prompt feature backed by openai. the growth chart bent vertical. twenty million sign-ups in eighteen months, mostly from people who wanted to see a generated deck once and then close the tab (close enough). in early 2024 the company raised eighty-one and a half million dollars at a three-hundred-million valuation, on the strength of those install numbers.

the revenue story never matched the install story. arr stayed under four million dollars through 2024. semafor reported in april '24 that tome was laying off staff to refocus on enterprise sales, and in october '24 the founders posted that they were stepping away from the slide product entirely to build something else. honestly, the official slide shutdown date was march 2025 (announced), then april 30 2025 (actual). everything not exported by midnight that day was deleted.

~ the case file ~

~ the climb ~

the september 2022 growth curve is the part of tome that became a slide in someone else's deck for the next year. before the ai pivot tome had maybe a few hundred thousand users, all of them in the figma-adjacent design crowd. after the pivot the curve went from a few hundred to one or two million per month, peaking in the spring of 2023 when generative-ai novelty was at its highest. tome's pitch deck started leading with that chart.

the founder talking points around the peak were ambitious. honestly, peiris and liriani positioned the product as the future of how presentations would get made, with humans giving outline-level direction and the model producing the actual slides. in fairness, the demos were credible. the generated decks looked competent on a phone screen for ten seconds. either way, the product never crossed the gap from interesting demo to thing people came back to weekly.

~ the slide ~

the conversion problem showed up early. tome was free, the paid tier added export-to-pdf and remove-watermark options, and most of the install base never hit a wall that made them want to pay. anyway, the unit economics of letting twenty million users run gpt-4 prompts for free were not workable. gpt-call costs dwarfed subscription revenue by a wide margin through 2023.

the market got crowded fast. gamma, beautiful.ai, decktopus, plus ai, slidegmm, presentations.ai, and a long tail of openai-wrapper slide tools all launched in the year after tome's pivot. each one pitched the same generate-deck-from-prompt loop. in those days. by mid 2024 the question of who would own the ai-presentation category looked like nobody, because the customers who would have paid for it just used powerpoint with copilot or claude in another tab. tome's growth flattened and the layoffs followed.

~ how it ended ~

the founders' october 2024 post was unusually direct for a san francisco startup goodbye. peiris wrote that the team had "made every effort to improve tome slides and turn it into a viable business but ultimately failed to find a sustainable path" (close enough). the path forward was that the team would build something new, in a different category, with the same investor base mostly intact.

that something new turned out to be lightfield, an ai-native crm for early-stage sales teams. the "tome" brand and url got sold to angellist, which rebranded a document-summarization product under the name. anyway, the new tome.app is a different product owned by a different company with no connection to the slide tool besides the four letters. anyone who searches for tome today lands on the angellist version.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »