Skip to main content

Stateless Chat

Stateless Chat allows your agent to send a single message to a model and receive a response — without tracking prior conversation history.

This is useful for:

  • Answering one-off user queries
  • Performing background lookups
  • Implementing batch processing workflows
  • Content summarization

Because no conversation state is maintained, each request is fully independent.

🔗 Endpoint​

POST /telexai/chat

Required Headers​

  • X-AGENT-API-KEY: Your API key

  • Model can be specified via:

    • X-Model header
    • model query param
    • model in request body

See Chat Overview for examples.

Scenario Examples​

One-off User Question​

A customer support agent needs to look up a product detail in real time:

Request:

{
  "messages": [
    {
      "role": "user",
      "content": "What is the return policy for orders over $500?"
    }
  ],
  "stream": false
}

Response:

{
  "status": "success",
  "status_code": 200,
  "message": "chat completed successfully",
  "data": {
    "Messages": {
      "role": "assistant",
      "content": "For orders over $500, the return period is extended to 45 days. Please ensure the item is in its original packaging."
    }
  }
}

Content Summarization​

An agent is tasked with summarizing a customer review for internal analysis:

Request:

{
  "messages": [
    {
      "role": "user",
      "content": "Summarize this review: 'I absolutely love this product! It exceeded my expectations in every way, and the customer service was fantastic. Highly recommend to anyone looking for quality and value.'"
    }
  ],
  "stream": false
}

Response:

{
  "status": "success",
  "status_code": 200,
  "message": "chat completed successfully",
  "data": {
    "Messages": {
      "role": "assistant",
      "content": "Positive review highlighting product quality, excellent service, and strong recommendation."
    }
  }
}

Usage Notes​

  • Stateless Chat is best for single-turn use cases
  • No context is preserved between calls — send full message in each request
  • Use in combination with dynamic model selection (Model Discovery)

Next​