generate() - RunAnywhere Documentation

const result = await RunAnywhere.llm.generate('Name three colours')
console.log(result.text)

Signature

One method takes either a prompt or a conversation:

generate(
  input: string | readonly ChatMessage[],
  options?: LlmOptions
): Promise<GenerationResult>

This differs from Swift and Kotlin, which have two overloads. On Web there is one method and a union input.

With options

const result = await RunAnywhere.llm.generate('Explain quantum computing', {
  model: 'qwen3-0.6b',
  maxOutputTokens: 200,
  temperature: 0.7,
})

console.log(result.text, result.tokensPerSecond)

Generation auto-loads what it needs, and downloads when options.model names a model that is absent.

LlmOptions

Field	Type	Default
`model`	`string`	none; uses the loaded model
`maxOutputTokens`	`number`	proto default
`temperature`	`number`	proto default
`topP`	`number`	proto default
`systemPrompt`	`string`	none
`stopSequences`	`string[]`	`[]`
`reasoning`	`ReasoningOptions`	none
`tools`	`ToolDefinition[]`	`[]`
`toolChoice`	`ToolChoice`	`auto`
`maxToolCalls`	`number`	`5`

Structured output

const result = await RunAnywhere.llm.generateStructured(
  'Extract the name and age: Dana is 34.',
  schema,
  'validationOnly'
)

Mode	Behaviour
`'validationOnly'`	Generate freely, then validate. The default.
`'repair'`	Validate, then retry once with a repair instruction.
`'constrained'`	Throws today.

'constrained' throws unsupportedCapability. It needs engine-level constrained decoding, which is not wired in yet.

Errors

Throws SDKException when no model can be loaded or generation fails. See Error handling.