Questions & answers

A sentence. Describe what you want — “a habit tracker with streaks and reminders” — and the builder plans it, asks anything it truly needs, and starts building. You watch progress live and get a running preview, not a mockup.

Websites and landing pages, full web apps with accounts and databases, 2D and 3D games, mobile apps (Android today; iOS on the roadmap), desktop apps for Windows/macOS/Linux, and browser extensions for Chrome and Firefox.

Yes — every build has a chat thread. Ask for changes in plain language and a new version is produced alongside the old one; every version stays downloadable and restorable.

Simple sites land in minutes. Apps with backends, games, and native builds take longer — native app packaging alone (real Android toolchain) can add 15–25 minutes. You can leave the page; the build continues server-side.

Yes. Builds run a verification loop — automated code review, a security audit, link/SEO checks, and accessibility testing — and the agents fix what the verifiers find before handing the build to you.

Yes — real engine rendering, not canvas fakes. Web games use a bundled three.js + physics stack; Unity Lab builds compile through the actual Unity 6 engine, and Blender Studio assets are authored headlessly in real Blender.

A public gallery of games our agents built end-to-end through Unity — code, 3D assets, audio — with build pages showing exactly what was made. See Unity Lab and Blender Studio.

Yes — music and sound effects are produced with sampled instruments (84 instruments plus ambience libraries), not beeps from an oscillator.

Android (.aab / .apk), Windows, macOS, and Linux desktop apps (Electron), and browser extensions for Chrome and Firefox. iOS/macOS App Store builds are on the roadmap.

Desktop apps are currently ad-hoc signed, not notarized by Apple, so the first launch needs right-click → Open. The install dialog on your build page shows the exact steps for your macOS version.

Yes — macOS builds are universal binaries.

Yes — the Ship team submits to Google Play, Chrome Web Store, and Firefox Add-ons using your own developer accounts. You connect credentials once; the agents prepare listings, privacy declarations, and upload the build. Store review itself is up to each store.

For store shipping, yes — the stores require the publisher to be you. Publishing to the web (your subdomain or your own domain) needs no accounts at all.

One click publishes to yourname.buildmidas.com. You can also deploy to your own server over SFTP, or point your own domain at the build.

Yes — dynamic apps (accounts, databases, APIs, even WebSocket multiplayer) publish to your subdomain with the backend hosted and managed by the platform.

Builds ship with robots.txt, sitemaps, per-page metadata, and pass an SEO verifier before completion. Post-ship, the Optimize team watches real Search Console data.

Five research agents sweep trends, social discussions, Q&A sites, and vertical data sources, then converge on written Opportunity Briefs — each one a concrete product idea with the evidence, the audience, and an MVP scope. Every brief has a “Build this” button.

Of course — Discovery is optional. Type your own prompt in the builder and skip straight to building.

Yes — like every team, it can run once, on demand, or continuously in scheduled autonomous mode, feeding fresh briefs to your queue.

A channel strategist picks the platforms worth your time for a given product, then per-platform writers draft genuinely native content — a Reddit post with a first-comment strategy, an X thread, a Quora answer that teaches first, a Medium article, LinkedIn and Facebook/Instagram posts.

Both, depending on platform. You can connect your own Reddit, X, or LinkedIn account and post drafts directly from the dashboard; for other platforms you copy the draft, post from your own account, and record the URL so traffic gets attributed.

No. Drafts are always shown to you first; nothing goes out on a connected account without your explicit action.

The Ads team plans campaigns and drafts ad creative & targeting. Money-touching steps stay manual by design — you approve budgets and launch in the ad platform.

Connect Google Search Console and Google Analytics and the Optimize team ingests real queries, clicks, and sessions daily, finds gaps (pages with impressions but poor click-through, content opportunities), and proposes or applies fixes.

Yes — store analytics snapshots (installs, listing performance) are collected daily for supported stores and joined with your web analytics.

Only in the modes you allow. It can run report-only, propose actions for your approval, or in autonomous mode apply and verify changes — with rollback.

They shouldn't: shared Analytics properties are always filtered by hostname, so a property collecting for several sites still reports only the connected domain. If numbers look wrong, check the property binding in My Domains.

Every unit of agent work — a build, a research run, a marketing draft, an optimization pass — consumes credits. Plans include a monthly allowance that resets each cycle; credit packs never expire and are consumed after the allowance.

The run pauses and tells you, rather than failing silently. Top up or wait for the next cycle and re-run.

Platform credentials power the agents by default. Bring-your-own-credential support is planned for teams who want their own accounts to do the work.

Charges are in USD. The interface shows an estimated local-currency figure for your region.

Yes — enough to build and publish something real before paying anything. No card required.

Unused credit packs within 14 days, first-charge-unused cycles, and platform-fault runs are all covered — see the Refund Policy.

You do — prompts, content, and the products the agents build in your workspace, including commercial use. See Terms §5.

Tenant isolation is enforced at the database layer with row-level security, and agent workloads run in isolated environments with restricted network access. Connected credentials are scoped to your account only.

Only power your own reports and optimization runs — never advertising, never other tenants. We adhere to Google's Limited Use policy; see Privacy §4.

Delete builds, domains, and credentials in-app anytime; for full account deletion, contact us from your account email and it's done within 30 days.

The dashboard and marketing pages run in 20 languages — switch with the globe icon in the top navigation. Agents reply in the language you write to them.

Each team supports three modes: run once, run manually whenever you ask, or scheduled autonomous mode where the loop keeps running — research feeding builds feeding shipping feeding optimization — until you stop it.

The public Showcase, Unity Lab, and Blender Studio — every entry built end-to-end by the agents.