Blog
Product updates, engineering notes, and ideas from Piskes.
The same audit that flags accessibility and SEO problems in the editor also runs as an API-key-authenticated endpoint — which means you can ask a real question: did this deploy make things worse?
Connections in website-builder work like the 'connected apps' page you already know from Google or GitHub — a list of what has access, and a revoke button that actually cuts it off.
AI website generators have a reputation for producing pages that all look the same, or pages that look like nothing in particular. The fix here wasn't a better prompt — it was taking invention out of the generator's hands.
The 500-template gallery isn't one thing — it's two production methods sitting behind one browsing experience, and knowing which is which changes what you should expect from a given template.
macca and mono got into the template gallery without ever passing through the reviewed files. In August 2026 we pulled them back out of the live database and fixed that — without changing anything a visitor sees.
Six pricing-model acronyms show up across the platform's campaign-type catalog, and most people outside ad buying couldn't reliably tell you the difference between them. Here's what each one actually charges you for.
Most SaaS dashboards default to filled-in sample data so the product looks alive before you've connected anything real. The ads app's dashboard doesn't — and that was a deliberate choice, not an oversight.
Every platform page in the ads app carries a 'Coming Soon — Integration In Progress' banner. Here's a direct answer to what that means and what's genuinely usable right now.
You don't need a live spend sync to start planning a multi-platform ad budget. The campaign-type catalog's pricing-model data is enough to build real numbers before you open a single ad account.
Process mining is normally a separate enterprise product line — Celonis, UiPath Process Mining — sold to ops teams analyzing human-driven processes. Workflow-worker ships a version of it for automations you've already built.
Instead of one CI pass/fail signal, Piskes enforces code quality across 17 apps and roughly 41 Workers with a large, registered, continuously growing set of checks that run before code can be committed or pushed at all.
An audit trail across 17 apps is a mechanics problem, not a feature checkbox — what's captured, what enforces coverage, and, just as importantly, what isn't there yet.
17 apps and roughly 41 Workers share one D1 database — here's specifically what happens if it goes down, and what's actually been tested versus what's designed but unproven.
Most all-in-one workspace platforms write payments as if Stripe is the only rail that matters. Piskes's payment layer supports MoMo and PayOS as first-class options, not an afterthought bolted onto a US-centric default.
A cross-app launcher solves a different design problem than a redesigned single-app nav — and it's worth being specific about which problem it actually solves, and which one it doesn't yet.
The consolidation argument has been made on opinion for years. Here's what the actual 2026 cost data says, what it doesn't say, and where a real all-in-one price structure lands relative to a typical fragmented stack.
Traditional SaaS margins ran 75-80% because serving another customer mostly meant another database row. AI-native software doesn't get that for free — every model call is a real, variable cost, and a platform's pricing should say so honestly.
Height was widely positioned as the most AI-native project management tool on the market — autonomous backlog grooming, bug triage, spec maintenance. It shut down anyway. That's worth sitting with before betting your workflow on one AI-native point product.
Even narrowly-scoped, single-purpose tools are re-architecting themselves around AI-agent orchestration in 2026 — evidence that starting from shared cross-app agent infrastructure, rather than adding agent features to one app at a time, was the more durable bet.
94% of organizations worry about vendor lock-in, and only 6% believe they could switch their primary vendor without material disruption. Here's what that number actually implies, and what's concretely built to address it versus what isn't.