A public log of how Rolo ships.

What shipped

• The product has been moving quickly since March but none of that velocity was visible to anyone outside the team. Invisibility at an early stage is a real credibility problem, and showing the actual pace of work is one of the most honest signals you can send.
• We chose a data-driven JSON approach rather than a CMS because the changelog needs to be maintainable by an AI agent on a weekly cadence. Hard-coded page arrays cannot be updated without touching route code, so a validated JSON source with a generation script was the right call.
• The retroactive backfill from March 2 was intentional. We wanted the full arc of what has been built to be readable in one place, not just the most recent few weeks. Most changelogs feel like they started last month.

What shipped

• Expanded mobile surfaces across dashboard, people, notes, capture, events, search, and profile in a single sprint rather than shipping one screen at a time and leaving the mobile experience half-finished for months.
• Improved auth flows and development ergonomics around signing in, making the production sign-in experience feel smooth and credible while reducing friction during active development.
• Built momentum on the companion mobile experience without compromising the product's visual identity. The same typography, color system, and interaction patterns carry through rather than diverging into a second design language.

What shipped

• Improved mobile headers, tap targets, safe area insets, and navigation behavior across the whole app. Mobile users notice the small things in a way desktop users often forgive.
• Cleaned up interaction bugs including edge cases in bottom navigation, modal layering issues, and focus management problems that broke keyboard input in certain flows.
• Made the mobile experience substantially closer to something people can trust and return to every day, which matters more for a relationship product than almost any other category.

What shipped

• Improved mention handling so that tagging a person in a note correctly links back to their full profile and history, rather than creating a disconnected text reference with no relational context.
• Made search and message drafting behavior more trustworthy by reducing the gap between what users intend when they type a query and what the system returns. That gap erodes confidence faster than almost anything else in a product that presents itself as intelligent.
• Focused on correctness over new screens, because a product that works reliably in the areas users already visit is more valuable than one with many features that feel slightly broken.

What shipped

• Advanced the product toward paid usage by wiring in billing infrastructure for the features that deliver the most value, deeper timeline access, rank-mode search, and fully assembled meeting prep.
• Improved environment stability, startup reliability, and health check behavior so the production deployment can fail fast and clearly instead of silently degrading in ways that are hard to diagnose.
• Handled the infrastructure work that would otherwise become a blocker mid-sprint, so the next few weeks of feature work could stay focused on product instead of stopping to wire up payment systems.

What shipped

• Added more structured in-app notification behavior so the product can surface time-sensitive context, like a relationship going quiet or a follow-up that is overdue, at the right moment.
• Improved how the product surfaces the next meaningful action for each contact, moving toward a model where Rolo actively helps you maintain relationships rather than passively storing data and waiting.
• Made the app feel more alive between major feature launches by investing in the connective tissue of the product, the small moments of guidance that give a tool genuine daily utility.

What shipped

• Reworked the marketing site by treating each feature section as a self-contained mini-application rather than a static illustration. The search showcase runs a real scoring function, the voice demo has real waveforms, because screenshots of a thinking product do not convey the thinking.
• Added interactive feature demos that let visitors experience the product's intelligence before signing up, including a search bar that scores contacts against a real dataset and a meeting prep flow built from actual contact history.
• Improved public-facing pages including the blog, product overview, features, compare, and audience-specific landing pages. The compare page in particular needed to explain what makes Rolo different without just listing feature checkboxes.

What shipped

• Locked in Coolify as our self-hosted deployment target on a VPS, giving us full control over the environment without the cost and constraints of managed platform hosting.
• Documented and planned around infrastructure explicitly so that product work could move faster later. When deployment is an afterthought it becomes a blocker at the worst possible moment.
• Invested in migration workflows, health checks, and environment management that would survive the product growing from a prototype to something real users depend on.

What shipped

• Pushed the notes experience closer to a real daily-use workflow where capturing something about a person or a meeting feels natural and fast, not like filling out a form.
• Improved settings and user controls so the app felt less like a prototype and more like something built with real user needs in mind. Settings are where users go when something feels wrong or when they want to make a tool feel like theirs.
• Focused on product usability rather than feature count, raising the floor so that existing surfaces felt genuinely trustworthy to use before moving on to the next thing.

What shipped

• Expanded the product around practical relationship workflows instead of generic CRM mechanics. Most relationship tools treat people as rows in a table and ask users to manage that table themselves.
• Improved the path from raw data to usable people context so that looking someone up gives you the full picture of your history with them, not just a sparse contact card.
• Set up the groundwork for smarter search, richer notes, and meeting prep features. The search system in particular needed a solid data layer before AI components could produce trustworthy results.

What shipped

• Moved beyond concept documents into implementation-ready workflows with real data behind them. This was the week Rolo stopped being an idea and started being something you could actually open and use.
• Began shaping the core experience around people, notes, and events, the three primitives everything else would depend on. Getting that data model right early prevented painful rework later when the schema would have fought the product.
• Created the structural foundation that later features could build on without reinventing their own patterns. Good early structure means engineers can add new capability without reverse-engineering the existing data layer each time.

What shipped

• Locked in the initial product direction around search, prep, and relationship memory. Three capabilities most CRMs gesture at but never actually solve well, which gave us a clear north star instead of getting pulled into feature sprawl.
• Started the repository and the planning system that now tracks every major product and engineering decision. Getting that infrastructure right early meant we could move fast on features without losing the reasoning behind past choices.
• Established the foundation for retroactive shipping updates so that the full arc of what gets built stays readable over time. Most early-stage products ship constantly but never document the story of it.