Twigg

Meet Twigg

Twigg provides a stateful API for applications that use language models. Developers create a chat once, then send each new message using its conversation identifier. The service stores earlier turns, prepares context, calls the selected model, and streams the response back with usage and cost information.

This moves conversation storage and context preparation out of the application's own servers. Developers can switch supported models between turns and control how much history a conversation carries. Twigg also handles model-specific message formatting and common request or streaming problems. It is intended for chat applications, assistants, and agents whose developers want to focus on their product while using a managed conversation layer.