integraciones-api

Connect the OpenAI API with Cerewro: GPT-4, embeddings and DALL-E from the chat

Use Cerewro to directly call the OpenAI API: completions with GPT-4o, embedding generation for semantic search, image creation with DALL-E 3 and audio transcription with Whisper. Combine models in custom pipelines.

22 de August de 2026

Equipo Cerewro

openai gpt api ia modelos embeddings fine-tuning

Connect the OpenAI API with Cerewro

Use Cerewro to directly call OpenAI's different services: completions with GPT-4o, embeddings for semantic search, images with DALL-E 3 and audio transcription with Whisper. Combine models in custom pipelines.

Service	Endpoint	Use
Chat Completions	`/v1/chat/completions`	GPT-4o: text, analysis, code
Embeddings	`/v1/embeddings`	Semantic search, clustering
Images	`/v1/images/generations`	DALL-E 3: generate images
Audio	`/v1/audio/transcriptions`	Whisper: audio to text

Transcribe meetings with Whisper

"Transcribe meeting recording C:\meetings\meeting-2026-09-02.mp3 using Whisper API in English. Then:
1. Generate meeting minutes with topics and decisions
2. List assigned tasks with owner and deadline
3. Save minutes as Word in C:\meetings\minutes\"

API costs: GPT-4o-mini costs ~$0.15/million input tokens. For medium-sized documents, costs are cents. Use GPT-4o-mini for classification/extraction tasks and GPT-4o only for maximum reasoning quality.