Connecting an AI provider to Maverick unlocks AI chat, automated project updates, and intelligent resource scheduling. Before you paste your first API key, here are five things worth understanding — they'll save you time and prevent common setup mistakes.

1. An AI provider is a company that runs AI models in the cloud

OpenAI, Anthropic, Google Gemini, and Microsoft Azure are all AI providers. They host powerful language models you access over the internet via an API. Maverick connects to these providers to power its AI chat and project-update features — you bring the account and key, Maverick handles the integration.

2. An AI model is the specific engine that processes your requests

Different models have different strengths — some are faster and cheaper, others are more accurate and capable. When you configure Maverick, you choose both the provider (the company) and the model (the specific AI engine). GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro are all models; OpenAI, Anthropic, and Google are the providers that host them.

3. API keys are how providers authenticate you

Instead of a username and password, providers issue API keys — long strings of characters that you paste into Maverick once. Every request Maverick sends to the AI is authenticated using your key, so usage and billing are tied directly to your account. Treat an API key like a password: copy it once into a secure location, then into Maverick.

4. Usage is billed by the token, not by the minute

Tokens are chunks of text — roughly four characters each. Providers charge a small amount per thousand tokens processed. Most project-management prompts cost only a few cents. Maverick shows you which model each resource is using, so you can track costs per person and right-size your model choices to each role's needs.

5. You can connect more than one provider at the same time

Maverick supports multiple AI providers simultaneously. You might use OpenAI for most users, Anthropic for executives who need higher accuracy, and a local model for sensitive projects. Each resource — person, machine, or workgroup — can have its own provider and model, giving you granular control over capability, cost, and data access.