Providers & Routing

DDTokens can distribute requests to available model resources through default routing. Default routing is recommended for most users.

Default Routing

When you use a model name directly, DDTokens selects a default available channel:

gpt-4o-mini

Default routing is suitable for:

• General conversations and text generation.
• Applications that need less configuration.
• Use cases that are not sensitive to provider differences.

Select a Channel

If DDTokens supports provider prefixes, you can pin an upstream provider with a format like:

openai/gpt-4o-mini
anthropic/claude-3-5-sonnet
google/gemini-1.5-flash

Available prefixes are determined by the model list in the dashboard.

When to Select a Channel

Consider selecting a channel when:

• You need a consistent response format.
• You have verified that a provider supports specific parameters better.
• You want to compare price, speed, or reliability across providers.

Notes

• Selecting a channel may change the price.
• Selecting a channel may reduce automatic failover capability.
• Different channels may support different parameters.
• Test with real production requests before launch.

If you are unsure, start with default routing.