Use OpenAI's GPT models to generate accurate, natural email drafts from your knowledge base.
OpenAI's GPT models offer industry-leading natural language generation. Relay lets you choose GPT as your AI engine for email responses.
#1
Most popular LLM provider
GPT-5
Latest model supported
95%
Draft accuracy rate
Link your Gmail or Outlook account in minutes.
Upload your FAQ, docs, and policies.
Relay reads incoming emails and drafts replies from your knowledge base.
Your team reviews AI drafts and sends with one click.
Incoming Email
Do you use OpenAI for generating responses? Can we choose which model to use?
AI-Generated Draft
Yes! OpenAI is one of our supported LLM providers, and you have full control over model selection. Here's how it works: **Supported OpenAI Models:** - GPT-5 — Most capable, best for complex technical responses - GPT-5 Mini — Great balance of quality and speed - GPT-5 Nano — Fastest, ideal for routine responses **How to Configure:** 1. Go to Settings > Agent > AI Model 2. Select 'OpenAI' as your provider 3. Choose your preferred model 4. Optionally use your own API key for direct billing **Key Features:** - All models draft from your knowledge base (not general internet knowledge) - Configurable response length, tone, and style - Model can be changed per agent — use different models for different mailboxes **API Key Options:** - Use Relay's included AI credits (simplest) - Bring your own OpenAI API key (for volume discounts or enterprise agreements) Would you like guidance on which model is best for your use case?
Access the latest OpenAI models for highest quality drafts.
Choose the right model for each mailbox or use case.
Use your OpenAI API key for direct billing and volume pricing.
GPT models draft from your knowledge base, not general knowledge.
Yes — bring your own API key for direct OpenAI billing. This is ideal for enterprises with volume agreements.
GPT-5 Mini offers the best balance for most support use cases. GPT-5 is better for complex technical responses.
Yes — change models anytime in your agent settings. No data loss or reconfiguration needed.
Start your free trial and see how Relay handles openai integration with AI-drafted responses.