Connecting AI model providers
Connect your own API keys from OpenAI, Anthropic, Google, and XAI to access AI models with dedicated rate limits and cost control.
Why connect providers?
Using your own API keys (BYOK) is optional. Consider connecting providers when you need:
Dedicated rate limits: No sharing with other Runtype users
Cost control: Direct billing from providers for separate cost tracking
Full model access: Use any model, including beta releases
High-volume production: Applications exceeding platform key execution limits
See Platform keys vs Bring Your Own Key (BYOK) for comparison.
Supported providers
OpenAI: GPT-4o, GPT-4, GPT-3.5, DALL-E
Anthropic: Claude (e.g. Sonnet, Opus, Haiku)
Google: Gemini (e.g. Pro, Flash)
XAI: Grok
See the full list in Available AI models.
Add a provider
You can add models two ways:
Model-first: Click Add Model to pick from the catalog; if the model requires an API key, you'll be prompted to enter it.
Provider-first: Configure a provider's API key in My Providers, then enable models from that provider.
To add a provider and enable models:
Go to Settings → Models.
Either click Add Model to add a model from the catalog (you'll be prompted for an API key if needed), or open the My Providers tab and click Configure next to a provider (OpenAI, Anthropic, Google, or XAI).
If configuring a provider: enter your API key from that provider and save.
Choose which models to enable (in My Models or when adding a model).
Click Save where applicable.
Getting API keys
OpenAI
Go to platform.openai.com.
Sign up or log in.
Navigate to API keys.
Click Create new secret key.
Copy the key (starts with
sk-).Paste it into Runtype Settings → Models (via My Providers → OpenAI → Configure, or when adding a model that uses OpenAI).
Anthropic
Go to console.anthropic.com.
Sign up or log in.
Navigate to API Keys.
Click Create Key.
Copy the key (starts with
sk-ant-).Paste it into Runtype Settings → Models (via My Providers → Anthropic → Configure, or when adding a model that uses Anthropic).
Go to makersuite.google.com/app/apikey or aistudio.google.com.
Sign in with your Google account.
Click Create API Key.
Copy the key.
Paste it into Runtype Settings → Models (via My Providers → Google → Configure, or when adding a model that uses Google).
XAI
Go to console.x.ai/keys.
Sign up for access.
Navigate to API keys.
Generate and copy the key.
Paste it into Runtype Settings → Models (via My Providers → XAI → Configure, or when adding a model that uses XAI).
API keys are sensitive credentials. Never share them publicly or commit them to version control. Store them securely.
Model selection
After adding a provider key, choose which models to enable:
In Settings → Models, find the provider (or open Add Model to see models by provider).
Check the models you want to use (or add them to My Models).
Click Save (or confirm) where applicable.
Enabled models appear in model dropdowns throughout Runtype (Flows, Agents, Playground).
Multiple providers
Connect multiple providers to access different model families:
OpenAI for GPT-4o and DALL-E
Anthropic for Claude's long context
Google for Gemini's multimodal capabilities
Switch between models in Flows and Agents as needed.
Cost tracking
When using BYOK:
Runtype shows estimated token usage
Actual billing happens directly with the provider
Check provider dashboards for precise costs
Set up billing alerts with providers to avoid surprise charges
See Usage data for Runtype's usage estimates.
Removing providers
Go to Settings → Models.
Open the My Providers tab and find the provider.
Click Manage keys (or equivalent), then remove the API key you no longer want.
Confirm removal when prompted.
Flows using that provider's models will fall back to platform keys (if available) or fail.
Best practices
Start with one provider: Add others as needed
Enable only models you'll use: Reduces clutter in dropdowns
Monitor provider billing: Set up alerts for spending
Rotate keys periodically: Security best practice
Test after adding: Verify keys work before deploying
Next steps
Managing AI models to configure model settings
Usage data to monitor costs
Using prompt steps to use connected models in Flows