Chat and models

A chat combines a task, its history, a provider model, project context, and the capabilities you allow. Use one chat per coherent outcome so its context and review remain understandable.

Connect a provider

In Settings, add a supported provider with your own account, subscription, or API credentials. Complete any provider authorization and verify that the connection is ready. Inference requests travel directly from the app to that provider; Maestrly has no hosted backend and does not resell model access. Provider authentication and retention follow the provider’s terms.

If authentication fails, check the credential, account status, region, and any organization or project selected at the provider.

Choose a model

Choose a model when creating a chat or in the chat's model control. Match the model to the work:

• use faster, lower-cost options for exploration and small edits;
• use more capable reasoning options for broad, ambiguous, or high-risk changes;
• check whether the model supports the tools, context size, or media the task requires.

The list can vary by provider account, plan, region, and current availability. A model shown by name is not an endorsement or a guarantee that your account can use it.

Write the first request

Include four things:

1. Outcome — what should be true when the work is done.
2. Scope — the project, files, or areas the agent may change.
3. Constraints — conventions, compatibility requirements, and actions to avoid.
4. Verification — tests, commands, or evidence expected before completion.

Attach only relevant context. Large, unrelated context consumes the provider's context window and can make results less focused.

Change models deliberately

Before switching models in an active chat, summarize unresolved decisions and current state. Confirm that the target model supports the required tools. After switching, check the agent's understanding before authorizing consequential actions.

Troubleshoot a request

• Unauthorized or forbidden: reconnect the provider and verify account, project, organization, or model permissions.
• Model unavailable: select another available model or check the provider's status and regional access.
• Rate or quota limit: wait, reduce parallel work, or adjust limits in the provider account.
• Context too large: remove unrelated attachments, split the task, or start a focused chat with a concise summary.
• Tool unsupported: choose a compatible model or remove the capability from the task.

Provider logs and billing dashboards are the source of truth for request status and charges. See Usage and costs.