OpenAI

Integrate OpenAI's powerful AI models for chat completion, embeddings, and speech-to-text. Build intelligent features with GPT-4, generate embeddings for semantic search, and transcribe audio with Whisper.

Setup

# .env
OPENAI_API_KEY=sk-...
OPENAI_ORGANIZATION=org-...

Chat completions

capabilityId: openai-chat
provider:
  type: openai
  operation: chatCompletion

inputs:
  messages:
    type: array
    items:
      type: object
      properties:
        role: string
        content: string
  model:
    type: string
    default: "gpt-4"
  temperature:
    type: number
    default: 0.7
    optional: true

outputs:
  content:
    type: string
  usage:
    type: object

Embeddings

capabilityId: openai-embeddings
provider:
  type: openai
  operation: createEmbedding

inputs:
  text:
    type: string
  model:
    type: string
    default: "text-embedding-3-small"

outputs:
  embedding:
    type: array
    items:
      type: number

Whisper (Speech-to-Text)

capabilityId: openai-transcribe
provider:
  type: openai
  operation: transcribe

inputs:
  audioFile:
    type: file
  language:
    type: string
    optional: true

outputs:
  text:
    type: string
  language:
    type: string

Best practices

  • Use streaming for real-time chat responses
  • Cache embeddings to reduce API costs
  • Implement rate limiting to avoid quota issues
  • Store conversation history for context
  • Monitor token usage and costs
Previous: Google CalendarNext: ElevenLabs