Create a Chatbot

Build an AI chatbot that answers questions based on your content. Add knowledge sources from URLs, sitemaps, or uploaded documents, then embed it on your site.

What you are building

A chatbot with RAG (retrieval-augmented generation) that reads your content, indexes it into a vector database, and answers user questions with source citations.

When to use this approach

• Customer support automation
• Documentation assistant
• Internal knowledge base search
• Website FAQ bot

If you need direct LLM API access without a knowledge base, use the LLM API instead.

What you need

• A WAYSCloud account
• Content to index (website URL, sitemap, or documents)

---

Step 1 — Create the bot

Dashboard

1. Open Services → AI & Machine Learning → Chatbot in the dashboard
2. Click Create Bot
3. Fill in name, purpose, and branding (colors, welcome message)
4. Select a plan
5. Add your first knowledge source (URL or upload)

API

curl -X POST https://api.wayscloud.services/v1/dashboard/chatbot/bots \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Support Bot",
    "slug": "support-bot",
    "language": "en",
    "purpose": "customer_support",
    "source_type": "url",
    "url": "https://docs.example.com",
    "display_name": "Support Assistant",
    "welcome_message": "How can I help you today?",
    "primary_color": "#2563eb",
    "selected_plan": "CHATBOT-BUSINESS"
  }'

Response:

{
  "id": "bot-abc123",
  "name": "Support Bot",
  "status": "draft",
  "public_token": "wayscloud_cb_abc123_full_token...",
  "source_count": 1,
  "created_at": "2025-12-15T10:00:00Z"
}

Save the public_token — it is only shown once.

---

Step 2 — Add knowledge sources

Add additional content for the bot to learn from:

Add a URL:

curl -X POST https://api.wayscloud.services/v1/dashboard/chatbot/bots/bot-abc123/sources \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "source_type": "sitemap",
    "url": "https://example.com/sitemap.xml",
    "title": "Main Website",
    "refresh_interval_hours": 24
  }'

Upload a document:

curl -X POST https://api.wayscloud.services/v1/dashboard/chatbot/bots/bot-abc123/sources/upload \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -F "file=@product-guide.pdf"

Sources are indexed automatically. Monitor progress:

curl https://api.wayscloud.services/v1/dashboard/chatbot/bots/bot-abc123/health \
  -H "Authorization: Bearer $JWT_TOKEN"

---

Step 3 — Test and activate

Test the bot before going live:

curl -X POST https://api.wayscloud.services/v1/dashboard/chatbot/bots/bot-abc123/test-chat \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"message": "What is your return policy?"}'

Response:

{
  "message": "Based on our documentation, the return policy allows...",
  "sources": [
    {"title": "Returns & Refunds", "url": "https://example.com/returns", "relevance": 0.95}
  ]
}

When satisfied, activate:

curl -X POST https://api.wayscloud.services/v1/dashboard/chatbot/bots/bot-abc123/activate \
  -H "Authorization: Bearer $JWT_TOKEN"

---

Step 4 — Next steps

• Embed on your site — Use the public token to add the chat widget
• Monitor analytics — GET /bots/{id}/analytics for session counts and feedback
• Tune behavior — Adjust temperature, max_tokens, and system_prompt
• Add guidance — Use help_text and avoid_text to steer responses
• Rotate token — POST /bots/{id}/rotate-token if the token is compromised

Related services

• Chatbot (CaaS) — full service documentation
• LLM API — direct language model access
• Object Storage — store knowledge base documents

Related guides

• Authentication — API key setup

API reference

• Chatbot API — all chatbot endpoints