Upgrade to Pro to get implementation-ready specs for every company, the full report library, and 5 on-demand report requests per month.
If you only have a few minutes to spare, here’s what investors, operators, and founders should know about reMail (W09).
reMail was a mobile email-search company founded in 2008 and accepted into Y Combinator's Winter 2009 batch. Its iPhone app delivered full-body search when Apple's own search mostly covered headers. The product changed architecture within months: version one indexed encrypted mail on reMail's servers, while version two moved the mailbox and index onto the phone after users objected to handing a startup their email credentials.[1][2][3]
Google acquired reMail in February 2010 and returned founder Gabor Cselle to the Gmail team.[4] The acquisition was an outcome, not evidence of commercial collapse. Yet the product's short life exposes a durable mechanism: privacy concerns can force an architecture change that also rewrites pricing, performance, and distribution.
Gabor Cselle entered reMail with unusually direct preparation. His biography says he earned an MS in computer science from ETH Zurich, worked as a Google software engineer, and became Xobni's vice president of engineering.[5] YC describes him as Xobni's first employee and says he later spent two and a half years at Google on Gmail and Android after the acquisition.[1] YC's company record names both Cselle and Einar Vollset as founders, while also listing a team size of one; the evidence does not clarify the operating split.[1]
The product attacked a specific gap in the first iPhone email experience. Apple's forthcoming iPhone OS 3.0 could search message headers, but not full bodies. Cselle argued that the phone's storage and processing limits made a complete local index difficult in 2009.[2] reMail's May beta therefore sent mailbox data to an optimized server, returned full-text results in seconds, suggested queries, grouped results into threads, and cached two weeks of mail plus previously searched messages for limited offline use.[2]
The architecture solved the hardware constraint but created a trust constraint. Users had to provide email credentials, and reMail stored encrypted mailbox data on its server. TechCrunch identified that arrangement as the largest adoption concern despite the company's promise not to inspect messages without permission.[2]
By August, reMail 2.0 had moved the full mailbox and index onto the iPhone. The app compressed 100,000 emails into about 500 MB, could resume an overnight initial download, and returned offline results in roughly one or two seconds.[3] The evidence set contains no verbatim founder quotations, so this report does not invent them. The observed product change speaks plainly enough: users valued search, but not on terms that transferred their inbox credentials to a young vendor.
The first reMail was a search service wrapped in an iPhone application. A user connected an inbox, reMail's server built a full-text index, and queries returned matching messages in seconds. Suggestions helped form searches, and matches appeared as threads. Users could reply or forward from results, but could not compose a new message. Accounts were capped at 10,000 messages.[2]
That version split the experience between fast remote search and a small offline cache. It planned to charge $3.99 per month after beta.[2] The model made economic sense for a server that continuously stored and indexed mail. It also asked users to trust reMail with credentials and encrypted mailbox copies.
Version two reversed the system boundary. The app downloaded the mailbox over IMAP, parsed MIME messages and attachments, compressed and stored content locally, and built its search index on the device. The first synchronization could run overnight and resume after interruption. Once complete, search worked offline and returned results in one or two seconds.[3] Cselle supplied a claim that this was about five times faster than Apple's search, though TechCrunch did not independently verify the benchmark.[3]
The client-only design also changed the product's economics. A recurring subscription became a one-time $4.99 purchase because reMail no longer carried the same indexing infrastructure.[3] After the acquisition, the local architecture let installed copies continue working without reMail's servers. The released code handled IMAP downloads, MIME parsing, attachments, and local storage, offering other email developers a documented starting point.[9]
reMail targeted iPhone users with large mail histories who needed to find information inside message bodies. Its early 10,000-message limit and later claim that 100,000 messages fit in about 500 MB show the product moving toward serious email users rather than casual inboxes.[2][3] Rackspace supplied the clearest distribution partnership by endorsing the app, simplifying configuration, and discounting account support for its email customers.[6]
No reliable figures for downloads, paid users, revenue, or addressable market appear in the evidence. Any market-size estimate would be invented. What can be established is that mobile email search was important enough for Apple to add header search in iPhone OS 3.0, for a specialist app to deepen it, and for Google to acquire the company within a year of launch.[2][4]
The decisive competitor was the platform. In 2009, Apple's search lacked full-body indexing, which gave reMail a clear wedge. Today Apple Mail searches addresses, subjects, message bodies, documents, and links, with Top Hits plus date and attachment refinements.[10] Gmail supports operators for senders, dates, attachments, labels, size, categories, and exact phrases, and eligible users can ask Gemini to find email.[11]
Apple Intelligence raised the baseline again in June 2024 with Priority Messages, inbox and thread summaries, and Smart Reply.[12] Faster keyword search is therefore no longer an independent product. A credible modern entrant must cross account boundaries, protect content from a new intermediary, and turn retrieved messages into useful work. Even then, Apple and Google can copy features inside their own ecosystems. The defensible space lies between providers, not inside one provider's mailbox.
reMail tested two prices that mirrored two architectures. The server-indexed beta planned a $3.99 monthly subscription, consistent with recurring storage and compute costs.[2] The client-only release charged a one-time $4.99 because the phone absorbed those costs.[3] Rackspace users could add an IMAP account for $0.99 instead of the usual $3.99, suggesting partner-led discounting as an acquisition tactic.[6]
YC and Gmail veterans Paul Buchheit and Sanjeev Singh funded the company, but no reliable total was found.[4] Revenue, conversion, retention, acquisition price, and paid-user counts were not disclosed. That makes it impossible to judge the standalone business. The pricing shift does show that privacy was not a narrow product objection: moving computation onto the device removed the natural basis for subscription revenue.
Version one's server index solved an authentic technical constraint. The 2009 iPhone lacked the storage and processing comfort for easy local full-text indexing, while Apple's own search stopped short of bodies.[2] Yet reMail's remedy required credentials and server-held mailbox data. That was precisely the information users were trying to search, and therefore the information they were least willing to hand to an unknown intermediary.
The attempted fix was unusually complete. reMail 2.0 did not add a privacy notice or a security badge. It eliminated the server from the search path, downloaded mail to the phone, indexed locally, and worked offline.[3] The change addressed the objection, but it also exchanged a $3.99 subscription for a $4.99 purchase and imposed an overnight initial download. Trust, gross margin, and onboarding were tied to the same architectural decision.
This is reMail's non-obvious mechanism: a privacy constraint can determine the business model. Server-side search supported recurring fees but created credential friction. Client-side search improved trust and removed infrastructure costs but weakened recurring revenue and moved setup pain onto the user. Neither version could optimize all four dimensions at once on 2009 hardware.
reMail entered through a missing platform feature. Apple was already closing part of that gap with iPhone OS 3.0, though its search remained shallower.[2] Google then acquired reMail and removed it from the App Store.[4][7] Existing users benefited from unlocked paid features and continued local operation, but the standalone product stopped acquiring customers.
The strongest counter-explanation is success: Google saw a useful product and hired an experienced email builder. Cselle's acquisition statement said the goal had been to reimagine mobile email, users had found it useful, and Google was the best place to continue work on communication and information sharing.[4] That account fits the evidence better than a distress narrative. The acquisition price remains unknown, and no source ties a later Gmail feature directly to reMail technology.
The evidence index does not preserve a verbatim acquisition quotation, so this section paraphrases the observed statement rather than manufacturing a quote. It also cannot establish whether Google bought product, code, talent, or some combination. Cselle's move to Gmail, the immediate App Store withdrawal, and the later code release suggest talent and domain expertise mattered at least as much as maintaining an independent app.[1][8]
The client-only pivot produced an unexpected benefit at shutdown. Because the app no longer depended on reMail servers, installed copies kept functioning. Releasing the client under Apache 2.0 exposed the difficult plumbing around IMAP, MIME, attachments, and local storage to other developers.[8][9]
That did not preserve reMail as a business, but it reduced user harm and extended the engineering work beyond the acquisition. Architecture determined exit quality just as it had determined trust and pricing.
Ready to rebuild reMail?
Implementation-ready specs, every report, and 5 on-demand requests each month.