Interface PromptRequestBaseBeta

Instruct the LLM how it should generate content. Be as specific and detailed as needed.

The LLM only has access to information in the instruction, plus the target schema.

String template with support for $variable from instructionParams.

param values for the string template, keys are the variable name, ie if the template has "$variable", one key must be "variable"

Examples

Constant

Shorthand

client.agent.action.prompt({
  instruction: 'Give the following topic:\n $topic \n ---\nReturns some facts about it',
  instructionParams: {
    topic: 'Grapefruit'
  },
})
Copy

Object-form

client.agent.action.prompt({
  instruction: 'Give the following topic:\n $topic \n ---\nReturns some facts about it.',
  instructionParams: {
    topic: {
      type: 'constant',
      value: 'Grapefruit'
    },
  },
})
Copy

Field

client.agent.action.prompt({
  instruction: 'Give the following field value:\n $pte \n ---\nGenerate keywords.',
  instructionParams: {
    pte: {
      type: 'field',
      path: ['pteField'],
      documentId: 'someSanityDocId'
    },
  },
})
Copy

Document

client.agent.action.prompt({
  json: true,
  instruction: 'Given the following document:$document\nCreate a JSON string[] array with keywords describing it.',
  instructionParams: {
    document: {
      type: 'document',
      documentId: 'someSanityDocId'
    },
  },
})
Copy

GROQ

client.agent.action.prompt({
  instruction: 'Return the best title amongst these: $titles.',
  instructionParams: {
    titles: {
      type: 'groq',
      query: '* [_type==$type].title',
      params: {type: 'article'}
    },
  },
})
Copy

Controls how much variance the instructions will run with.

Value must be in the range [0, 1] (inclusive).

Default: 0.3