EZ-Bot.ai | Developer API Documentation

How It Works

1

Sign up at ez-bot.ai and go to the Developer tab in your dashboard

2

Generate your API key and copy the AI prompt

3

Paste the prompt into Claude Code, ChatGPT, or any AI agent

4

Your AI reads GET /schema, discovers all capabilities, and builds your bot

5

It uploads your content, configures design, tests answers, and confirms everything works

6

Add the embed code to your website — your chatbot is live

Authentication

All requests require a Bearer token in the Authorization header.

# Every request needs this header Authorization: Bearer cbp_your_api_key_here

Generate your API key from the Developer tab in the dashboard. Rate limit: 120 requests/minute.

Quick Start (Python)

# Full bot setup in Python — or let your AI agent run these calls import requests API = "https://chat.ez-bot.ai/api/v2" HEADERS = {"Authorization": "Bearer YOUR_API_KEY", "Content-Type": "application/json"} # 1. Check what's available schema = requests.get(f"{API}/schema", headers=HEADERS).json() # 2. Configure the bot requests.patch(f"{API}/config", headers=HEADERS, json={ "bot_name": "Acme Support", "primary_color": "#2563eb", "welcome_message": "Hi! How can I help you today?", "allowed_domain": "acme.com", "system_prompt": "You are a helpful support agent for Acme Corp.", "starter_questions": ["What do you sell?", "Return policy?", "Contact support"], }) # 3. Upload knowledge base content requests.post(f"{API}/knowledge/documents", headers=HEADERS, json={ "filename": "faq.md", "content": "# FAQ\n\n## Return Policy\nWe offer 30-day returns on all items...", }) # 4. Add custom responses for instant answers requests.post(f"{API}/knowledge/responses/bulk", headers=HEADERS, json={ "responses": [ {"trigger_phrases": "return policy, refund", "response": "We offer 30-day returns. Visit acme.com/returns"}, {"trigger_phrases": "pricing, cost, price", "response": "Plans start at $19/mo. See acme.com/pricing"}, ] }) # 5. Test the bot result = requests.post(f"{API}/test/batch", headers=HEADERS, json={ "tests": [ {"question": "What is your return policy?", "expected_keywords": ["30", "return"]}, {"question": "How much does it cost?", "expected_keywords": ["$19"]}, ] }).json() print(f"Tests: {result['passed']}/{result['total']} passed") # 6. Confirm ready health = requests.get(f"{API}/health", headers=HEADERS).json() print(f"Status: {health['status']}") # "ready"

Endpoint Reference

Discovery & Health

Method	Endpoint	Description
GET	`/schema`	Full API schema — AI agents call this first
GET	`/health`	Bot readiness (LLM up, KB chunks, config status)

Account

Method	Endpoint	Description
GET	`/account`	Account info, plan, limits, usage stats
PATCH	`/account`	Update company name or email
PUT	`/account/password`	Change password

Widget Configuration

Method	Endpoint	Description
GET	`/config`	Read current widget configuration
PATCH	`/config`	Partial update — only send fields to change
PUT	`/config`	Replace entire configuration

Knowledge Base

Method	Endpoint	Description
GET	`/knowledge/documents`	List documents with chunk counts
POST	`/knowledge/documents`	Upload file or raw text content
DELETE	`/knowledge/documents/<id>`	Delete document + embeddings
GET	`/knowledge/responses`	List custom keyword responses
POST	`/knowledge/responses`	Create single response
POST	`/knowledge/responses/bulk`	Bulk create/replace responses
PUT	`/knowledge/responses/<id>`	Update response
DELETE	`/knowledge/responses/<id>`	Delete response
POST	`/knowledge/crawl`	Crawl a website for KB
GET	`/knowledge/crawl/status`	Poll crawl progress
GET	`/knowledge/sync/<doc_id>`	Get auto-sync config
PUT	`/knowledge/sync/<doc_id>`	Update sync config

Testing

Method	Endpoint	Description
POST	`/test/chat`	Test message with debug info (no usage tracking)
POST	`/test/batch`	Batch test with expected keywords — pass/fail
POST	`/test/webhook`	Test webhook URL
POST	`/test/smtp`	Test SMTP config

Analytics & Conversations

Method	Endpoint	Description
GET	`/analytics`	Full dashboard analytics
GET	`/analytics/conversations`	Conversation list (paginated)
GET	`/analytics/conversations/<id>`	Conversation messages
GET	`/analytics/leads`	Lead list (paginated)
GET	`/analytics/handoffs`	Handoff requests
POST	`/analytics/handoffs/<id>/resolve`	Resolve handoff

Live Chat Inbox

Method	Endpoint	Description
GET	`/inbox`	List inbox sessions
GET	`/inbox/<session_id>/messages`	Get session messages
POST	`/inbox/<session_id>/reply`	Send operator reply

Integrations & Actions (Pro/Business)

Method	Endpoint	Description
GET	`/connectors`	Available connectors (Shopify, Stripe, Lemon Squeezy, etc.)
GET	`/integrations`	List integrations
POST	`/integrations`	Create integration (manual or connector)
PUT	`/integrations/<id>`	Update integration
DELETE	`/integrations/<id>`	Delete integration
GET	`/integrations/<id>/actions`	List actions
POST	`/integrations/<id>/actions`	Create action with slots
PUT	`/actions/<id>`	Update action
DELETE	`/actions/<id>`	Delete action
POST	`/actions/toggle`	Enable/disable action bot

Quality Assurance

Method	Endpoint	Description
POST	`/quality/run`	Start automated QA test
GET	`/quality/runs`	List QA runs
GET	`/quality/runs/<id>`	QA run results

Key Concepts

Knowledge Base Upload

Two ways to add knowledge:

Raw text (recommended for AI agents): Send JSON with content and filename fields. Great for AI-generated content.

File upload: Send multipart/form-data with a file field. Accepts .txt, .pdf, .md, .csv.

Custom Responses

Keyword-triggered instant responses bypass the AI entirely. Perfect for FAQ answers with direct links.

trigger_phrases: comma-separated keyword groups. All words in a group must appear in the visitor's message. Most specific match wins.

Batch Testing

The /test/batch endpoint lets your AI verify the bot works correctly. Send questions with expected keywords — get a pass/fail for each.

Test results don't count toward usage limits. Run as many tests as you need.

Pre-Built Connectors

Connect Shopify, Stripe, Zendesk, and more with just API credentials. Call GET /connectors to see available types and required fields, then POST /integrations with the credentials.

Built for AI Teams