Twigg

Twigg is a stateful LLM interface service designed for developers. Ordinary large-model APIs are stateless; each request requires the retransmission of the entire conversation history, and tasks such as context management, message truncation, and model switching must be implemented manually. Twigg eliminates these complexities by handling these functions internally.

The usage is very simple: create a conversation, and then simply send the command “Next Event”. Twigg is responsible for storing and assembling the relevant context on the server side, and it will automatically trim the data according to the specifications of the target model. Since the conversation data is stored with Twigg rather than with any specific model vendor, it is possible to switch models halfway through the conversation – without being restricted by any particular vendor.

We provide OpenAPI documentation as well as an llms.txt file specifically designed for AI systems. These resources are ideal for development teams working on chat products, Agent applications, or those that need to seamlessly switch between different models.