@gotong/llm-openai

OpenAI LlmProvider for Gotong. Plug into @gotong/llm's LlmAgent.

Install

pnpm add @gotong/llm-openai openai
# openai is a peer dependency — install the version you want

Use

import { LlmAgent } from '@gotong/llm'
import { OpenAIProvider } from '@gotong/llm-openai'

const provider = new OpenAIProvider({
  // apiKey: '...'         // defaults to process.env.OPENAI_API_KEY
  // defaultModel: '...'   // defaults to 'gpt-4o-mini'
})

const reviewer = new LlmAgent({
  id: 'reviewer',
  capabilities: ['review'],
  provider,
  system: 'You return one revision suggestion.',
})

Translation rules

• LlmRequest.system is hoisted to a leading { role: 'system', content } message.
• Output cap uses max_completion_tokens (the modern OpenAI field), not the deprecated max_tokens.
• finish_reason: 'stop' → 'end_turn'; 'length' → 'max_tokens'; anything else ('tool_calls', 'content_filter', …) → 'error'.
• usage.prompt_tokens / completion_tokens → inputTokens / outputTokens.

License

MIT