Skip to content

Chat Agent Guide

Overview

Section titled “Overview”

Chat agents provide text-based conversational AI with full conversation threading, history, and transcript export. They’re ideal for customer support, documentation assistants, FAQ systems, and lead qualification.

Quick Start

Section titled “Quick Start”
  1. Get a text model — Check available models with GET /models?type=TEXT
  2. Create a chat agent — POST /chat-agents with your model ID and system prompt
  3. Send a message — POST /chat/send with your agent ID and message
  4. Continue the conversation — Include conversationId in subsequent messages

Configuration

Section titled “Configuration”

System Prompt

Section titled “System Prompt”

Define your agent’s personality, knowledge, and behavior. The system prompt is sent with every conversation turn to maintain consistent agent behavior.

Model Parameters

Section titled “Model Parameters”
  • Temperature — Controls response creativity (0.0 = focused, 1.0 = creative)
  • Max Tokens — Limits response length for cost control
  • Provider — Choose between Gemini, OpenAI, or Grok text models

Supported Text Models

Section titled “Supported Text Models”
ProviderModelsBest For
OpenAIGPT-4o, GPT-4o MiniPremium quality, complex reasoning
GeminiGemini 1.5 Pro, Gemini 1.5 FlashCost-effective, general purpose
GrokGrok BetaReal-time information, web search

Conversation Management

Section titled “Conversation Management”

Viewing Conversations

Section titled “Viewing Conversations”
  • GET /chat/conversations — List all conversations for an agent
  • GET /chat/conversations/:id — Get full conversation with message history

Ending Conversations

Section titled “Ending Conversations”
  • POST /chat/conversations/:id/end — Mark a conversation as ended
  • DELETE /chat/conversations/:id — Delete a conversation entirely

Exporting Transcripts

Section titled “Exporting Transcripts”
  • GET /chat/conversations/:id/transcript — Export as plain text

Token Tracking

Section titled “Token Tracking”

Bizway tracks token usage per conversation for cost optimization. Monitor your usage patterns to:

  • Right-size your max_tokens setting
  • Choose the most cost-effective provider for each use case
  • Identify conversations that consume disproportionate resources

Use Cases

Section titled “Use Cases”
  • Customer Support — Handle common questions and route complex issues to humans
  • Documentation Assistants — Help users find answers in your knowledge base
  • Sales Qualification — Qualify leads through conversational questionnaires
  • Interactive Help — Guide users through onboarding or troubleshooting flows