Multi-model AI assistant: choose a supported provider
HeyCC separates the bot you create from the AI service that processes its requests. You can configure and choose a model service supported by the installed version while remaining responsible for that provider's account, terms, privacy policy and fees.
Verified model-choice boundary
- A bot's role, name and appearance are separate from its selected model service.
- Relevant messages and attachments go to the configured AI endpoint.
- Provider quality, speed, availability, limits, terms and fees can differ.
- HeyCC does not claim to support every model, endpoint or provider feature.
What changing the model can affect
A model choice can change response style, reasoning quality, latency, context limits, availability and usage cost. A provider may also handle data under different terms. Compare those differences using the same low-risk task instead of assuming the largest or newest model is automatically best for every bot.
What changing the model does not change
Switching a supported provider does not automatically rewrite the bot's purpose, remove the need for tool permissions or make every operation safe to run without review. Model selection and action approval are separate decisions: one chooses who processes the request, while the other controls what the app may do.
How to compare model services in HeyCC
- Choose one small, non-sensitive task with a clear expected result.
- Run the same input through two supported configurations without changing the bot's instructions.
- Compare usefulness, factual errors, latency and any provider-reported usage.
- Review each provider's data policy, account limits and billing before using private material.
- Keep the configuration that fits the task, then repeat the test when the provider or installed app version changes.
Bring your own AI key for a supported provider
When the installed HeyCC version offers an API-key configuration for a supported provider, that credential authorizes requests to the provider; it does not make the model free, local or offline. Keep credentials private, enter them only through the app's provider configuration, and verify the provider account that will be billed.
Data flow and privacy
Relevant messages and attachments needed for an AI response are sent to the selected endpoint. Device-side tools or iCloud app-data sync are different paths and do not mean model requests stay on the device. Read the HeyCC privacy and data-flow guide and the selected provider's current policy before sending sensitive information.
Limits and user control
- Available providers and authentication methods can change by build, platform, account and region.
- Provider outages, model retirement, rate limits, billing and data practices are outside HeyCC's direct control.
- A configured endpoint can still reject a model name or unsupported request format.
- Test with low-risk content before relying on a new provider for important work.
Multi-model AI assistant FAQ
Can I choose the AI model service used by HeyCC?
Yes. HeyCC lets the user configure and select among model services supported by the installed version. Availability can change by build, platform, account and region.
Does HeyCC support every AI model and provider?
No. HeyCC supports specific configurations and does not claim compatibility with every model, endpoint or provider feature.
Are AI provider fees included with HeyCC?
Do not assume they are included. Provider accounts, limits, terms and usage fees vary. Check the selected provider and current configuration before sending a request.
Does using an API key make AI requests local?
No. An API key authorizes requests to that provider. Relevant messages and attachments are still sent to the configured AI endpoint.
Is choosing a model the same as approving an action?
No. Model selection chooses the service that processes a request. Tool permissions and action approvals are a separate control layer.