RVF LLMInstall

Summary

Generate audio content using a selected provider and model with configurable message and option settings.

Inputs

• provider (required) — AI provider loaded from RVF LLM API; only providers that support the selected resource are shown.
• model (required) — Model selected for generation, filtered by provider and resource type.
• messages (required) — Message definitions; each message includes a role (system/user/assistant) and content.
• options — Generation options such as temperature, max_tokens, top_p, simplify_output, stream, frequency_penalty, and presence_penalty.
• builtInTools — Optional tool toggles (webSearch, fileSearch, codeInterpreter) that can be enabled for the request.

Output shape

array of objects each containing a 'json' field with the formatted response

The array may include error objects {error: "..."} when continueOnFail is enabled.

Examples

Example 1: Generate a short poem using default settings

provider: "OpenAI"; model: "gpt-3.5"; messages: [{role: "user", content: "Write a haiku about sunrise"}]; options: {temperature: 0.7, maxTokens: 500}; builtInTools: {}

Example 2: Stream a response with temperature 1.0 and enable web search

provider: "Custom"; model: "custom-model"; messages: [{role: "user", content: "Summarize recent news"}]; options: {temperature: 1.0, maxTokens: 1500, stream: true, builtInTools: {webSearch: true}}; simplifyOutput: false

Example 3: Generate an audio clip using a specific provider

provider: "ElevenLabs"; model: "voice-1"; messages: [{role: "user", content: "Create a 5‑second greeting"}]; options: {temperature: 0.5, maxTokens: 300}; builtInTools: {}