RVF LLMInstall

Summary

Generate text using a selected AI provider and model

Inputs

• provider (required) — AI provider loaded from RVF LLM API; only providers supporting the selected resource are shown.
• model (required) — Model loaded from RVF LLM API, filtered by provider and resource type.
• options — Model configuration options including temperature, maxTokens, topP, simplifyOutput, stream, frequencyPenalty, presencePenalty.
• builtInTools — Optional built‑in tools such as webSearch, fileSearch, or codeInterpreter.

Output shape

a list of JSON objects representing the generated response

Each item contains the formatted output; if simplifyOutput is true only the main message text is returned, otherwise the full raw response. Errors appear as {"error": "..."} within the same structure.

Examples

Example 1: Generate a short poem about sunrise

provider: "openai", model: "gpt-4", options: {temperature: 0.8, maxTokens: 150}, builtInTools: {webSearch: true}

Example 2: Summarize a research article with streamed output

provider: "cohere", model: "cohere-regular-s2.1", options: {temperature: 0.5, maxTokens: 500, stream: true}, builtInTools: {}

Example 3: Translate a paragraph using built‑in code interpreter

provider: "anthropic", model: "claude-3-opus", options: {temperature: 0.3, maxTokens: 300}, builtInTools: {codeInterpreter: true}