LoginDaddyTempail alternative: copy, wait, leave
Tempail is a fine standard-issue temp mail page. Address, inbox, done. This post is not about what it does wrong — it is about what the category keeps adding around a job that has three verbs.
Copy.
The page hands you an address. You copy it. Any control beyond "copy" here is a feature looking for a user. Tempail's copy button works. Ours works. The verb is solved industry-wide and nobody gets points for it anymore.
Wait.
The message arrives or it doesn't. Waiting is the product — the page is a chair. Mail that lands in seconds vs minutes is a property of the sender's queue, not of the chair's upholstery. Judge waiting rooms by retention (ours: about two days; the box: up to three), not by furniture.
Leave.
The underrated verb. A new address should end the old box's visibility — on this page it does; the old mailbox is purged on rotation. Leaving should also mean the page forgets you: no account existed, so there is nothing to delete. That is the whole manifesto. Three verbs, one page.
Tempail does the job. If you want the job with fewer nouns around it: open the inbox. All notes.
Page weight as a feature
Minimalism is measurable. Count what loads before the address appears: HTML payload, stylesheets, scripts, fonts, ad slots and their auction libraries, trackers, each with DNS, TLS, and download time on the user's actual connection. On fiber, a heavy page costs a shrug; on a congested mobile network in a parking lot outside the store whose Wi-Fi gate needs the signup code, it costs the job. Temp mail is disproportionately used in exactly those degraded conditions — traveling, tethering, borrowing connectivity — because that is when throwaway addresses are needed most urgently. A page that renders its one address in a single small HTML response with deferred everything-else degrades gracefully where a framework-heavy page with five ad auctions times out. This is not an aesthetic argument but a reliability one: every kilobyte is a bet on the user's bandwidth, and minimal pages win the bet everywhere. Our page keeps the critical path tiny (address, buttons, list) and loads the rest after — not from design taste but because the address must survive bad networks. Measure any temp page by time-to-address on a throttled connection; the results reorder every ranking made on fiber.
Privacy minimalism: data never collected
The strongest privacy claim a page can make is architectural: there is nothing to leak because nothing was kept. No accounts means no credential database. No analytics means no behavioral profiles. Short retention with scheduled purges means the breach window is days, not years. Each absence removes a breach category, a subpoena target, and a retention-policy argument. Contrast with feature-rich pages that necessarily hold more: preferences, histories, API keys, billing records — each legitimate, each a liability. Minimalism here is not marketing but blast-radius engineering: the product's privacy posture equals what it refrains from building. Users should apply this lens to every free tool: list what it stores, ask what each stored item costs you if exfiltrated, and prefer the tool whose list is shortest for your job. The three verbs survive this audit trivially — copy keeps nothing, wait keeps mail briefly, leave deletes. That is the manifesto's technical content, and it fits in one paragraph because the architecture fits in one page.
The three-verb API
Think of the verbs as an interface contract, the way engineers spec APIs. Copy: the page MUST present a complete, valid address within seconds of load, selectable in one action, with no configuration required. Wait: the page MUST display arriving mail without refresh acrobatics, preserve it for a stated window (ours: about two days), and show status honestly (waiting vs failed vs expired). Leave: the page MUST offer rotation that purges the old mailbox, MUST NOT retain account-shaped residue, and SHOULD forget the session on disconnect. Any control beyond these verbs needs justification against the contract: does a domain menu serve copy, wait, or leave? Do statistics serve any verb, or the operator's curiosity? This lens evaluates features without taste arguments — a control either serves a verb or it is clutter with a constituency. Apply it to our page too: the Turnstile check serves wait (keeping the box abuse-free enough to receive mail), the new-address button serves leave, and everything else is suspect. Three verbs, one page, zero alibis.
No-JS fallbacks: the page without the script
Disable JavaScript and most modern pages become blank rectangles — including, honestly, ours past the static shell. But the degradation gradient matters: a temp-mail page should fail toward readability (address visible in HTML, messages listed server-rendered) rather than toward a spinner. Progressive enhancement is the old discipline with new relevance: core content in HTML, interactivity layered on top, every feature asking whether the job survives its absence. Copy works without JS (select the text). Reading works if the list renders server-side. Only live-updating truly needs script, and polling is a convenience over refresh, not a requirement. Users running script blockers, text browsers, or assistive tech that chokes on framework soup experience the fallback as the product — and in low-connectivity regions, the "fallback" is the only version that loads at all. Minimal pages get this mostly free: less script means less to fail. Audit any temp page with scripts off; what remains is what the builder actually respects.
Font and asset budgets: system stacks win
Webfonts are the heaviest non-functional payload on most pages: hundreds of kilobytes, render-blocking, flash-of-invisible-text while loading. A temp-mail page has no brand equity worth that weight — system font stacks render instantly, match the user's OS, and cost zero bytes. Same logic extends to every asset class: icon fonts replaced by inline SVG, frameworks replaced by vanilla script where the interaction is three buttons, images omitted where CSS suffices. The budget discipline is simple: every asset must justify itself against time-to-address on a throttled connection, and decorative assets never survive that audit. Our page carries fonts and framework CSS from a shared site shell — a genuine compromise for visual consistency across pages, openly acknowledged as weight the manifesto would otherwise cut. Perfect minimalism would inline everything; real products balance. The honest metric stays time-to-address, and every release should measure it rather than assume it.
The copy button: a micro-interaction audit
Copy looks like one line of code and behaves like a compatibility matrix. The modern clipboard API is asynchronous, permission-gated, and unavailable in insecure contexts — a correct implementation tries it, handles denial gracefully, and falls back to the legacy selection-and-exec-command path for older browsers. Feedback needs three states, not one: idle, copied (transient confirmation, then revert), and failed (with the address left selected for manual copying). Edge cases multiply: clipboard permissions denied at browser level, focus requirements on some mobile browsers, formatting that copies invisible characters along with the address. Test the matrix or inherit its bug reports. Our copy control implements exactly this ladder — modern API first, fallback second, manual selection always available — because a copy button that silently fails is worse than no button: the user pastes nothing into the form and blames the signup. Micro-interactions earn trust in inverse proportion to their size; copy is the smallest interaction on the page and the one the entire job depends on.
Status lines vs spinners: honest loading states
Loading indicators encode a theory of waiting. Spinners say "something is happening" without saying what — appropriate for sub-second operations, dishonest past a few seconds because they imply progress where none is measurable. Status lines say what state the system is in ("waiting for mail," "connected," "retrying") — informative at any duration because state is always knowable even when progress is not. The rule: show state always, progress only when measurable, and elapsed time passively so users can calibrate their own patience. Our kicker status line follows this — connection state and mailbox state as text, no spinners anywhere, because mail arrival has no progress bar and pretending otherwise is theater. Spinner abuse has a cost beyond honesty: users who cannot distinguish "working" from "stuck" file support tickets for normal waits and abandon broken flows that look normal. Every loading state should answer three questions — what is happening, is it normal, what do I do — and spinners answer none of them. Text does.
Tempail does the job. If you want the job with fewer nouns around it: open the inbox. All notes.
Take the three-verb audit beyond temp mail and it keeps working. Password managers: generate, fill, lock — anything beyond the verbs is either sync infrastructure (justified) or engagement design (suspect). Note apps: capture, find, export. Read-it-later: save, read, archive. Each category has products bloated past their verbs and minimal ones holding the line; neither camp wins universally, but the audit names exactly what you pay (in attention, money, or data) for every control past the core. Apply it to our own page quarterly and honestly — the day a fourth verb earns its place, add it deliberately with a changelog note, not by accretion. Minimalism maintained by audit stays minimal; minimalism maintained by taste drifts. The manifesto is a procedure, not a personality.
Open inbox