Frequently asked questions

ModelCrew is a multi-model AI workspace where you define AI teammates with roles and instructions, connect supported model providers, and coordinate work in projects. It is an orchestration layer, not a new foundation model.

ChatGPT and Claude offer polished consumer experiences. ModelCrew is designed for a different workflow: multiple persistent roles, potentially using models from different providers, working within the same project and coordination process.

No. ModelCrew accesses supported providers’ models through APIs. It does not reproduce every feature in ChatGPT, Claude.ai or other consumer applications.

ModelCrew supports delegation and collaboration workflows in which teammates perform tasks and a Project Lead can review and synthesize contributions. The exact workflow depends on the configured crew and available capabilities.

The teammate’s configured role and instructions are separate from its assigned model. Changing that model preserves the configured role, though results and capabilities may differ across models.

ModelCrew stores teammate configuration, including role and instructions. This is not the same as unlimited native memory in the underlying model; the application provides relevant context to model calls.

No. Models from different providers remain separate. ModelCrew can provide relevant project and conversation information to model calls, subject to context limits.

Choose based on task fit, quality, latency, cost, features and provider availability. Model rankings change; no provider is automatically the best for every type of work.

Bring Your Own Key means connecting your own provider API credentials so ModelCrew can use the models through those providers. API billing is separate from consumer subscriptions.

No. Consumer subscriptions and provider API billing are generally separate. You can keep a consumer subscription while also paying for API usage, depending on how you use both products.

No. Some task/model combinations can be cheaper, but orchestration across multiple teammates may use more model calls and raise total usage. Costs depend on the workflow.

ModelCrew has been built around connections to supported providers including OpenAI, Anthropic, Google, DeepSeek, Moonshot/Kimi and OpenRouter. Availability depends on the current release and your configured provider access.

No. API keys are sensitive credentials. Add them only through the app’s designated credential-connection process, and follow the relevant provider’s security guidance.

Do not assume that. Security and data-handling guarantees should be verified against the current product documentation and terms before processing sensitive content. Never interpret a BYOK setup as a privacy guarantee.

ModelCrew is designed for a human-in-the-loop approach with work review and approval controls. Specific workflows and available controls can evolve during beta.

ModelCrew is in beta. Use the Request access link in the website navigation to express interest. Access and availability may change as the beta progresses.