Frequently asked questions.

Maestra is an AI agent and a cloud platform. The agent composes business software from a conversation; the platform runs it, with a database, sign-in, audit trail, and an API included in every app. Everything it builds is standard code you own.

Companies lease dozens of tools that each fit their work a little badly, while their people quietly build replacements on personal accounts. Maestra turns that building into owned, governed software on one foundation.

Software the business depends on ends up scattered across laptops and side accounts, with no sign-in policy, no audit trail, and no owner once its builder moves on.

AI agents made building software cheap. Running it safely is still expensive. The gap between the two is where most internal tools stall or become a liability.

As a platform decision rather than a tool decision. One vendor review covers every app your people build on it, and every app inherits the same controls.

Usually one team starts with a real problem, ships an app in days, and the organization connects its directory once that app is in daily use. From there, builders in other teams start on the same foundation.

No. People describe what they need in plain language. Engineers who prefer to work locally can clone the repository and push like any other project.

Every app carries a score: a readable document describing what it does, for whom, and why. It is what the agent builds from, what reviewers read, and what keeps the app understandable after its builder leaves.

A typed relational database, vector search, managed sign-in, file storage, email in and out, scheduled jobs, webhooks, an HTTP API and MCP server, monitoring, and an append-only audit log.

Yes, through its integration catalog, inbound webhooks, and plain HTTP. Apps can call your internal APIs and your systems can call theirs.

Yes. Every app is standard TypeScript in a git repository your organization owns and can export at any time.

You keep the code and the data. Apps are built on an open framework and can be redeployed on your own infrastructure.

Hundreds of models across text, image, video, and voice, from every major lab and the leading open-weight families. You choose per agent and per app.

A security review packet, data-processing terms, a pre-filled questionnaire, and a penetration-test summary, all available on request during procurement.

Data is encrypted in transit and at rest, isolated per tenant, and only reachable through authenticated platform endpoints. Access is logged.

In the region you choose for your workspace: United States, European Union, Canada, or Australia.

Organization plans include a written uptime commitment and a support channel with named contacts.

The app belongs to the organization, not the person. Ownership transfers in one step, and the score explains the app to whoever inherits it.

Dedicated deployments are available for organizations with specific requirements. Talk to us about yours.

Directory-based roles, per-person model access, spend limits, data residency, audit export, and a single view of every app, owner, and permission.

Yes. The controls are defaults, not projects. A small team can govern hundreds of apps because every app is governed the same way.

Internal tools and workflows mostly: trackers, portals, approval flows, dashboards, intake forms, and customer-facing apps that sit close to operations.

Apps are responsive web applications that work well on phones. Native builds are on the roadmap.

Better models make the agent better. The platform, the part that runs, secures, and governs the software, matters more as more software gets built.

A flat $99 a month for the platform ($79 billed annually), plus AI tokens at the providers’ own per-token rates with no markup.

Either from a balance you fund, which never expires, or directly by the providers if you bring your own keys.

The product agent that works with you from idea to production and keeps working after launch.

Coding, design, roadmap, QA, architecture, and research specialists, conducted by Maestra itself.

It keeps the roadmap, writes the docs, watches production, fixes bugs, runs experiments, and drafts marketing.

A roadmap the agent maintains from how the app is used. One click starts building the next item.

Start a trial, build one real app, then contact sales to connect your directory and set organization-wide controls.

A first working version usually takes hours, not weeks. Most of the time after that is refinement with the people who will use it.