Story Settings API

Most story settings related to text generation can be accessed and modified through the API. This includes memory, system prompt, author’s note, prefill, and generation parameters. Scripts can both read the current settings and update them as needed. Updating these settings requires the script to have the storyEdit permission.

Memory

Memory is a persistent text field that is always included in the generation context. It typically contains important information about characters, plot points, and world details that should always be available to the model.

Getting Memory

To retrieve the current memory text, use the api.v1.memory.get function:

let memory = await api.v1.memory.get();
api.v1.log(`Current memory: ${memory}`);

Setting Memory

To update the memory text, use the api.v1.memory.set function:

await api.v1.memory.set('Character: Maria - A brave knight seeking redemption.');

Author’s Note

The author’s note (A/N) is inserted near the end of the context, just before the most recent story content, as a system message.

Getting Author’s Note

To retrieve the current author’s note and its settings, use the api.v1.an.get function:

let an = await api.v1.an.get();
api.v1.log(`Author's Note: ${an}`);

Setting Author’s Note

To update the author’s note, use the api.v1.an.set function:

await api.v1.an.set('Please please please write good.');

System Prompt

The system prompt provides instructions to the model about how to generate text. It’s placed in a system message at the start of the generation context.

Getting System Prompt

To retrieve the current system prompt, use the api.v1.systemPrompt.get function:

let prompt = await api.v1.systemPrompt.get();
api.v1.log(`Current system prompt: ${prompt}`);

Setting System Prompt

To update the system prompt, use the api.v1.systemPrompt.set function:

await api.v1.systemPrompt.set('Write good please.');

If the system prompt is set to an empty string the default system prompt will be used.

Prefill

Prefill is added to the start of the most recent “assistant” message in context. It can be used to strongly influence model responses.

Getting Prefill

To retrieve the current prefill text, use the api.v1.prefill.get function:

let prefill = await api.v1.prefill.get();
api.v1.log(`Current prefill: ${prefill}`);

Setting Prefill

To update the prefill text, use the api.v1.prefill.set function:

await api.v1.prefill.set('Here\'s a list of delicious baked goods:');

Generation Parameters

Generation parameters control the behavior of the text generation model, including temperature, token limits, and other sampling settings.

Getting Generation Parameters

To retrieve the current generation parameters, use the api.v1.generationParameters.get function:

let params = await api.v1.generationParameters.get();
api.v1.log(`Temperature: ${params.temperature}`);
api.v1.log(`Max tokens: ${params.max_tokens}`);

Updating Generation Parameters

To update generation parameters, use the api.v1.generationParameters.update function:

await api.v1.generationParameters.update({
  temperature: 0.7,
  max_tokens: 150,
  top_p: 0.9,
  top_k: 50
});

You only need to provide the parameters you want to change; other parameters remain unchanged. See the API Reference for a complete list of generation parameters.