landing/ollin-pro.html — a product-page preview, not a deployed site/ doc, so it isn't linked here; this page describes it in prose. The distribution + first-dollar strategy it serves is the oll.in launch & distribution kit →, and the architecture that will one day power the agent is the Agent Seam plan → (drive oll.am over MCP, not a new backend). Read the launch kit for where to post and what to charge; read this for why the page is built the way it is.
Every incumbent makes you do the robot's work. oll.in is the honest inverse — the agent does it, you approve the send.
◆ The one-sentence positioning
oll.in is an agent that finds roles that fit you, writes each Swiss application (photo, CV, cover letter, interview prep), and keeps your LinkedIn/X presence warm so recruiters come to you — and you review and send every one. It is deliberately built for the everyday professional in Switzerland & DACH — the nurse, the teacher, the tradesperson — the person for whom "paste your CV into ChatGPT for every job" is a nightly grind, not a dev workflow. That audience choice drives every design decision below.
01What oll.in actually is — and who it refuses to be for
A landing converts only if the first ten seconds name the reader's pain and the product's shape at once. oll.in's hero does both — and its audience choice is the strategy.
The hero headline is "The job hunt that runs itself while you work and sleep." The sub-line is the whole thesis in one breath: "Stop pasting your CV into ChatGPT for every job. Your agent finds and writes them — you just review and send." Two CTAs — Join the waitlist and See it work — and a trust row (you approve every send · cancel anytime · made for real jobs, not just tech) that pre-empts the three objections a non-technical reader has before they scroll.
The last of those — "not just tech" — is the deliberate line in the sand. Every auto-apply tool on the market is US-centric, English-first, and built for software engineers. oll.in's wedge is the opposite: the nurse in Bern, the teacher in Zürich, the plumber in Luzern — great at their job, allergic to job-hunting. The page repeats this everywhere (the hero faces, the who-it's-for grid, the walkthroughs) because the audience is the differentiation. A landing that tried to serve everyone would blur it.
⚠️ The agent boundary is a product decision, stated up front
The agent prepares; the human sends. This isn't a disclaimer bolted on at the bottom — it's in the hero trust row, its own "You send. Always." band under the steps, and the JSON-LD FAQ. It never auto-applies, never mass-applies, never runs a bot on any job platform. That boundary is both an honesty line and a moat: "apply to 100 jobs an hour" is the hostile, ToS-violating half that burns the user's own account — oll.in owns the safe, high-value half (find + tailor + prepare) and hands the click to a human on purpose.
02The nine-section conversion architecture
The page was decluttered to nine sections, each doing one job in the persuasion sequence: name the pain → contrast the enemy → show the mechanism → widen the value → prove it's for you → make it concrete → price it → humanise it → ask. Nothing is decorative.
The rhythm is classic conversion, but the content is disciplined to the audience: no jargon, concrete francs, real faces, and a single ask. The three red steps are the signature moments — the parts a reader remembers.
03The three signature moments
Most landing pages are text with buttons. oll.in earns belief with three concrete artifacts that make the abstract ("an agent works for you") tangible — each one honest by construction.
1 · The agent-at-work activity feed
The hero's right column is a mocked dashboard that shows the agent doing the work — a live-styled feed the reader can imagine waking up to. It's the fastest way to answer "what does an agent even do?" without a paragraph of explanation.
Note the last row: even the demo ends at "Review", never "Applied". The honesty boundary is baked into the hero art itself.
2 · The before/after Bewerbungsfoto — the founder's own photo
The deliverables section shows a professional-photo before/after. The critical craft choice: it uses the founder's own generated headshot, honestly labelled "A real example — the founder, both looks made with the photo tool. Not a customer testimonial." A stock before/after would be a lie; a real one from a customer they don't have yet would be a fabrication. Using Samuel's own face is the only honest way to show the result — and it doubles as proof he uses his own tool.
This is also where oll.in connects to the platform's existing foto-service headshot capability — the one product on Core that already has a guest-checkout path. The photo is the concrete, shippable-today slice of the agent's promise.
3 · The real profession photos
The who-it's-for grid uses real photographs of the actual professions — a nurse, a classroom, a workbench, an office desk, a café. For an audience that has never seen a product "for people like me", recognition is the conversion. The photos are honestly disclosed as stock (see §04), used for illustration — the faces are not presented as customers.
04The honesty architecture — the through-line and the moat
This is the point of the whole page. oll.in's honesty is not compliance decoration bolted to the footer — it is the positioning. In a category full of "apply to 1,000 jobs while you sleep" over-promises, being the one tool that tells the truth is the differentiation a nervous, non-technical, DACH audience actually buys.
| The temptation | What oll.in does instead |
|---|---|
| Fake testimonials with 5-star faces and "I got 3 offers!" | Labelled illustrative scenarios. Sofia/Marco/Luka are badged example; the copy says "what they'd wake up to", not "what they got". |
| "Auto-applies to hundreds of jobs for you" | The agent prepares; you review, approve and send. Never auto-applies, mass-applies, or runs a bot — stated in the hero, the steps, the FAQ, and the fine print. |
| Big permanent-looking prices, or a fake countdown timer | Prices are launch targets shown with a genuine limited-time founding discount; "may change"; the waitlist doesn't charge you. |
| "Swiss-hosted · your data stays in Switzerland" | No data-residency or hosting-location claim. "Swiss format" is explicitly scoped to CV layout conventions, nothing more. |
| Stock faces passed off as happy customers | Full disclosure: the before/after is the founder's own photo; every other image (professions grid, avatars, hero faces) is stock, for illustration — not a real customer. |
| "Grows your LinkedIn automatically" (a bot) | Official LinkedIn/X sign-in, drafts you approve. Scouting uses legitimate public listings and official APIs only. |
✦ Why honesty is the moat, not the tax
The auto-apply category is a graveyard of over-promises — which has trained this exact audience to distrust. A nurse who's been burned by a "1,000 applications" tool reads oll.in's "you approve every send · this is a real example, not a testimonial · prices may change" and feels something no competitor gives her: relief. The honesty is what makes the product believable enough to try. It converts because it's true, not despite the cost of being true. This is the same ethics line that governs the rest of oll.am — an honest tool, never a bypass.
05EN/DE — the whole reading path, both languages
The audience is DACH. A German speaker who hits an English wall bounces. So the page carries a full EN/DE language toggle — and it's engineered to cover the entire persuasion path, not just the nav.
How the toggle is built
✓ Auto-detects German browsers — a German-locale visitor lands in German; a DE/EN button in the nav flips it, and the choice persists.
✓ Every reading-path element is translated — an I18N_SEL selector list covers the eyebrow, hero headline + sub, all section heads and leads, the who-it's-for note, plan descriptions, CTA buttons, the founder quote, the price footnote and the promo banner. The conversion argument survives the switch intact.
✓ Rich-text aware — translations that carry emphasis (the underlined "arbeitest und schläfst", the bold "du prüfst nur und schickst ab") are marked rich:true so the German keeps the same typographic stress the English has.
Small tell of the craft: even the founding-offer banner and the honest price footnote are translated — because the moments where trust is won or lost (the ask, the small print) are exactly the ones a reader will scrutinise in their own language.
06The founding-price lifecycle
Pricing on an early-access page is a lifecycle problem, not a number. oll.in frames three plans around a single honest idea: hold a founding price now, pay when it opens.
The design choices that keep it honest: a 50% limited-time founding discount shown as a struck target price (the reader sees exactly what they save and that the target may move); the Job Hunt plan is framed as a 2-month start, billed monthly, cancel once hired — matching the real shape of a job search, not a trap subscription; the CHF 15 photo is a founding-price reservation, not a live charge; and a footnote states plainly that joining the waitlist doesn't charge you. The pricing sells the future without pretending the present is more than it is.
How the ask actually resolves
The waitlist form POSTs to Core's /api/waitlist with an email + source tag, and degrades to a plain mailto:hello@oll.in fallback if that fails. So the single conversion event — an email captured — works even before any billing is wired. First value is a real, low-friction signal, not a paywall.
07The craft under the surface
The details a casual read misses — the ones that make the page feel finished, findable and trustworthy.
Findability & sharing
✓ OG + Twitter share card (og:image 1200×630, en_CH with a de_CH alternate) so a shared link renders as a real card in a WhatsApp/LinkedIn preview — the channels this audience actually shares in.
✓ JSON-LD SEO — Organization, WebSite, Service and a FAQPage graph that answers the four real questions ("does it apply for me?", "is it live?", "how much?", "who's it for?") in structured data, honestly.
Feel & resilience
✓ Reveal-on-scroll with a safety timeout — sections fade in as you scroll, but a fallback timer force-reveals everything so the page can never get stuck invisible if the IntersectionObserver misfires. Motion is also disabled under prefers-reduced-motion.
✓ Dark mode — a full dark palette, auto-detected from the OS and toggleable, persisted to localStorage.
✓ One honest accent system — a warm red (#E23A2E) brand with amber/green/blue supporting roles; note oll.in runs its own product-brand palette (Inter + Fraunces), distinct from these oll.am docs (Playfair + Source Serif) — a product landing, not a build-log page.
⚠️ Status — this is a design preview, not a deployed page
The landing lives at landing/ollin-pro.html in the working tree. It is not a deployed site/ artifact, and oll.in is early access — opening autumn 2026. Nothing here charges anyone yet: the photo is a founding-price reservation and the plans are waitlist reservations. Per the ship-vs-build line, oll.in sits behind the current first-dollar path (CiteBible / the live money paths) — this landing is the ready-when-it's-time asset, captured now so the design decision and its honesty are on the record.
landing/ollin-pro.html) is a complete, decluttered nine-section conversion page for the Swiss/DACH job-application agent: an agent-at-work hero, the enemy contrast, Scout/Match/Prepare/Track, Career-SEO "get found", the professions grid, the illustrative Sofia/Marco/Luka walkthroughs, founding pricing, the founder, and the waitlist close — with a full EN/DE toggle, OG/JSON-LD SEO, dark mode, and an honesty architecture (no fake testimonials, agent-prepares/human-sends, launch-target prices, no "Swiss-hosted" claim, full stock-vs-founder photo disclosure) as the positioning moat. The distribution plan is the launch kit; the future agent architecture is the Agent Seam plan.