Build2 distinct publishers3 min readUpdated
Paid Relay workspaces stop at 11:59 p.m. Pacific on September 14th, and stored credentials are deleted with them. Jacob Bank takes the approval-checkpoint idea into Google Chrome.
The Engineer · Build desk

Compiled by The EngineerSomething wrong?How this is made
Relay, the no-code automation product that positioned itself between Zapier's trigger-based integrations and collaborative work-management tools such as Asana, is shutting down, and founder Jacob Bank is returning to Google to lead product and developer relations for Chrome, with some Relay employees joining him [1][14]. For anyone who put a recurring business process on it, the news that matters is the calendar, not the hire.
The wind-down started on July 16th, when Relay disabled new signups and upgrades and canceled paid subscriptions [3]. Free users got 30 days to export, with accounts scheduled for deletion at 11:59 p.m. Pacific on August 15th [4]. Paying customers got a 60-day transition ending at 11:59 p.m. Pacific on September 14th [5], which is 60 calendar days from the July 16th announcement [22]. Annual subscribers are to receive prorated refunds for unused billing periods, and each paid workspace was given an extra 25,000 workflow steps and 10,000 AI credits per month during the transition [6][7].
That extra quota is a bridge, not a migration path. Exports cover workflows, sequences and MCP servers as JSON files, plus prompts, run histories and tables [8]. What does not come out is the part that takes the longest to rebuild: Relay says accounts, stored credentials, authentication tokens and unexported data will be permanently deleted after each transition window [9]. Every connected app has to be reauthorized by hand wherever the workflow lands next, and the JSON cannot do that for you. Relay's service is disappearing rather than continuing under Google, while Bank and some colleagues move into Chrome [10], so there is no hosted fallback to drift into.
Bank's landing spot is specific. According to his LinkedIn profile, cited by mezha.net, he becomes vice president of product for Google Chrome, responsible for Chrome's product and developer relations teams [11]. In a post on X he called Chrome "a perfect place to collaborate with agents" and said he plans to help users get work done with AI in the browser [12]. He is not new to the building: he left a Stanford AI Lab Ph.D. program to co-found Timeful, which Google acquired in 2015, then spent a little over six years at Google in product leadership roles across Gmail, Google Calendar and Google Chat before leaving in July 2021 to build Relay [13].
The design he is carrying over is the interesting part. At Relay's October 2023 public launch, its AI steps summarized text, classified incoming email, extracted information and drafted content, each paired with human approval checkpoints so a user could inspect output before a message went out or a business record changed [15]. Bank argued publicly that fully autonomous agents often waste tokens and produce less predictable results when a process should follow the same plan every time [16]. Tasks could be assigned to a person or a role, letting a run halt for an approval, a missing field or a judgment call, across onboarding, meeting prep, support triage, document drafting and finance work [17].
The financing explains the shape of the outcome. Khosla Ventures led a $5 million round in 2022 with Neo, BoxGroup, SV Angel and angels; Andreessen Horowitz led a further $3.1 million announced with the 2023 launch [18][19]. Relay's own company page claims $8.2 million raised against $8.1 million announced, a $100,000 gap, and no valuation was disclosed [20][21]. Small totals, capable team, product retired.
Watch whether Chrome ships cross-app agent work with Relay's approval checkpoints intact, or without them. mezha.net notes Chrome already uses AI as a secondary assistant and frames Relay's automation experience as reinforcement for that direction amid Gemini's rollout across Google products [23]. And watch September 14th: after it, exported JSON is all that remains [8][9].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Relay is shutting down and founder Jacob Bank is returning to Google to lead product and developer relations for Chrome, according to TechCrunch's August 17th report; some Relay employees will join him.
Free Relay accounts closed on August 15th, while paying customers retain access through September 14th.
Relay began its shutdown process on July 16th, when it disabled new signups and upgrades and canceled paid subscriptions.
Free users had 30 days to export their work before their accounts were scheduled for deletion at 11:59 p.m. Pacific on August 15th.
Paying customers received a 60-day transition period ending at 11:59 p.m. Pacific on September 14th.
Relay's service will disappear rather than operate under Google, while Bank and some colleagues move into Chrome.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Dated shutdown mechanics, single-publisher depth
The spine — closure, staged cutoffs, July announcement, Bank's Chrome role and title — is carried by two independent publishers both crediting TechCrunch, and the operational specifics (times, refunds, credits, export formats, deletion scope) are concrete and internally consistent. Depth is nonetheless single-sourced to runtimewire.com, Relay itself is not quoted directly, no primary Relay notice or Google statement is included, and the sources disagree on Relay's launch year while the company's own total-raise figure does not reconcile with announced rounds.
Lifecycle events documented, usage undisclosed
Adoption here is measurable only as the shutdown's execution: five dated lifecycle events exist (signup freeze, free-tier deletion, transition credits and refunds, the scheduled paid cutoff, the Chrome appointment), which is real evidence that the wind-down is underway rather than announced. But no source discloses Relay workspace counts, revenue, retained customers, or how many employees move to Chrome, and Chrome-side integration of Relay's approach is unreported, so the installed base being retired is entirely unquantified.
Mildly overstated strategic upside
The factual reporting is restrained and well-dated, so the gap is small. It is positive rather than zero because the strategic framing runs ahead of evidence: one publisher suggests Relay's experience could strengthen Chrome's Gemini direction, and both foreground continuity between Relay's approval-checkpoint design and Chrome's agent work, while the sources report no acquisition, no asset transfer, no headcount, and no Google statement about the team's remit inside Chrome's roadmap.
Vendor-shaped wind-down narrative
Most substantive detail traces to parties with an interest in how the closure reads: Relay's own shutdown notices and company page (including the unreconciled $8.2 million total), and Bank's X post framing Chrome as the ideal next venue. Google is not quoted and no customer or investor voice appears, so the softening elements — refunds, extra credits, orderly export paths — are self-reported. Offsetting this, the deadlines and deletion scope are commitments that expose the vendor rather than flatter it, and one publisher explicitly flags the funding discrepancy and the absence of acquisition terms.
Solid on the what, thin on the scale
Two independent publishers agree on the closure, the cutoff dates and the Chrome appointment, and the operational detail is specific enough to act on, which supports fairly high confidence in the timeline and the data-deletion consequence. Confidence is held below the eighties because everything beyond the spine rests on one publisher restating vendor material, the launch-year discrepancy is unresolved, and the scale of affected customers and the Chrome-side outcome are entirely unreported.
product
Relay's shutdown hands Chrome a product boss and its customers a September 14 deadline2 distinct publishers
product
Rippling shipped the product it was sued over the moment both suits vanished1 distinct publisher
invest
Etched's $10.3B mark prices a non-Nvidia inference bet at ten times booked orders1 distinct publisher
invest
Tiny corp wants Etched's numbers. Jane Street led $700M at $21B without publishing any1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
mezha.net
1 article · August 17, 2026
runtimewire.com
1 article · August 17, 2026