Platform / Models & Agents

Every model, one SDK, no keys to juggle.

Every Maestra app is written against one SDK that carries hundreds of models across text, image, video, and voice, an agent runtime, and a large integration catalog. Builders call a model the way they call a method. The platform handles keys, routing, metering, and limits.

01

One call, any model

Swapping models is a one-word change. Nothing else in the app moves.

const summary = await ai.generate({ model: "frontier/reasoning", // or any of hundreds input: ticket.thread, schema: Summary, });
02

Every modality

Model catalog · by modality
Text & reasoning180+
Image60+
Speech & audio40+
Video25+
Embeddings20+
Counts are illustrative and change as labs release new models.
03

An agent runtime built in

Apps can run agents that plan, call the app’s own methods as tools, browse, and hand work back to a person. The runtime handles memory, retries, and timeouts, and every step lands in the audit log.

04

Access and spend, governed

Per-person model access
MemberFrontierOpen weightsImageMonthly limit
DWDana Whitfield AllowedAllowedAllowed$400
MRMarcus Reyes AllowedAllowed—$150
BCBea Coleman —Allowed—$40
Admins decide who can use which models, and how much.
At cost
Model usage is billed at the providers’ own rates, with no markup.
Bring your own keys
Or connect your own provider accounts and be billed directly.

Start building on Maestra.

Everything you just read is standard in every app, running from the first deploy.

Start building →
← Back to the Platform