Office Hours #11 · Open Mic Q&AEvery system,
one board.

Press Publish, then either go through the mobile app publishing flow (second tab), or simply navigate to your published app URL on your phone browser. On first visit, both iPhone and Android will show an "Install" button that adds it to your home screen as a PWA.

Two approaches. First, from your very first prompt, tell it "build this app mobile first, desktop second." Second, use AI Controls: click the cog icon bottom-left of the prompt window → AI Controls → Custom Instructions → add "only build a mobile app from a design point of view."

Type in chat: "We're going to do a big mobile UX overhaul. I want you to change the visuals, layout, structure — everything to do with just mobile. Leave desktop alone." Then iterate until it's right.

Sam doesn't know but called it "an absolutely fantastic question." He'll check if it's been logged as feedback and submit it if not.

The App Store won't allow any mention of Stripe. You need separate flows based on device identifiers — hide the Stripe component when viewing from a wrapped mobile app, show the native in-app purchase flow instead. For subscriptions, Revenue Cat is the recommended middleware — it unifies all subscription providers into one system.

Not currently through those specific channels, but Base44 runs frequent community events (vibe sprints, one-prompt challenges) that reward coupons and credits. Plus "something may be occurring this weekend."

Open one tab with your settings showing integration credits, open another tab as a user. Use every feature for 5 minutes and check how many credits it consumed. Then calculate cost per credit to determine pricing.

In-app purchases (one-time) should already work. In-app subscriptions are being worked on but no timeline available.

No, this doesn't work on annual plans currently. But something is in the pipeline to resolve this — Sam wouldn't give further details.

Paste that exact question into your Base44 chat and it will implement it. For more control, add admin backend controls for sensitivity levels and flagged words.

Currently yes, but it's all about to change. The new RLS system will have AI proactively suggesting security settings. For now: run the security scan, apply all rules, then test as a user. Deploy in phases: 5 family users → 10 real users → 30 from a mailing list → public.

Not yet. Sam checked the docs during the session and confirmed it's not currently HIPAA compliant, though it was being looked at.

The enterprise plan offers customisable rate limits and data hosting in the location of your choice (data residency). Pricing isn't publicly advertised and may vary per company. Sam recommended initiating via the "Talk to Sales" button. Enterprise also includes dedicated account management and phone support.

There are no storage limits at all. The main constraint is rate limiting, which enterprise plans can customise. There isn't a hard ceiling — it depends on your plan and architecture.

5,000 is the hard-set limit for data requesting/putting. Two workarounds: enterprise plan (customisable limits) or batching — slow down the process, implement a queueing system, make it process in smaller chunks.

No hard restriction on user count. The soft restriction is rate limiting at scale. Mitigations: caching, batching requests, using timers instead of live polling, and compartmentalising data. You could technically get 10,000 users overnight with no issue if the architecture handles it.

You can add up to 350 custom domains to your app via Dashboard > Domains. However, per-user custom domains are not currently supported. Sam offered to raise it internally.

Several approaches depending on scale. Small: contain everything in one app with organisation-level data separation. Larger: clone apps from a white-label template. Most scalable: use git sync to maintain a template, push to cloned instances. Sam noted this could fill a three-hour session.

Not out of the box. Post your app in the Discord showcase channel for natural community feedback. Also consider creating a Super Agent as your dedicated tester — it's a useful thought experiment and practical tool.

You can upload more than just text and MD — PDFs are now supported, among others. If there's a specific format you need that isn't supported, submit feedback.

Yes to everything. Payments, user registration, and tournament management are all standard out-of-the-box features. Use Plan mode to describe your requirements in plain language. For a custom domain, you need the Builder plan ($20–40/month).

If your app uses backend functions (payment providers, APIs, etc.), yes — downgrading below the plan that includes backend functions will break those features. The website itself would still exist, but payment processing would stop working.

Have the customer log into the app. Go to Dashboard > Users, change their account from user to admin, then transfer ownership to them. The customer needs any paid plan to take over.

That's just how AI works — you'll always get more than you asked for. Take a screenshot, draw crosses on every page you don't want, and tell it to remove them. Pre-prompt outside the platform, use discuss mode, and always use screenshots with drawn arrows.

The main con is AI aversion — many people have strong negative feelings about AI, especially in creative fields. Sam's advice: be transparent, tell them it's Base44, be a proud advocate. "Tell them Base44. Be a proud advocate of it."

Yes, especially in the early days — Sam vetted almost every prompt through ChatGPT. Now he prefers Discuss mode within Base44 since it has the app's context. But using external AI as a mentor/validator is absolutely recommended.

No. A screenshot circulated showing what appeared to be native dark mode, but after investigation it was determined to be fake. Tom's unofficial Chrome extension is the only dark mode option. There's a 10-month-old feature request with low traction.

It's a staged rollout — it's been released to a percentage of users and is gradually expanding. If you don't have it yet, you will soon.

Not for the dashboard code editor directly, but you can achieve this now by git syncing and controlling files at the git level.