🤖 Stylus Chatbot API

The Stylus Chatbot API allows any project or system to interact with an AI assistant capable of answering questions based on the official documentation of Stylus, the WASM smart contract framework for Arbitrum.

This API provides a simple HTTP interface to send prompts to a selected open-source language model and receive structured, cleaned responses.

---

🌐 Public Endpoint

POST https://stylus-demo.duckdns.org/api/stylus-chat

---

📥 Request

Headers
Content-Type: application/json

Body (JSON):

{
  "model": "deepseek-r1:7b",
  "prompt": "How do I deploy a Stylus contract?"
}

Field	Type	Required	Description
model	string	✅ Yes	The model to use for inference (see list below)
prompt	string	✅ Yes	The question or instruction sent to the chatbot

---

📤 Response

{
  "response": "To deploy a Stylus contract, you must first..."
}

Field	Type	Description
response	string	The AI-generated output, cleaned

---

✅ Supported Models

You can specify any of the following model names in the model field:

• deepseek-r1:7b
• deepseek-r1:14b
• llama3.1:8b
• qwen2.5:32b

⚠️ Model names must be written exactly as shown. Larger models may have longer response times.

---

🧪 Example with curl

curl -X POST https://stylus-demo.duckdns.org/api/stylus-chat \
  -H "Content-Type: application/json" \
  -d '{"model": "llama3.1:8b", "prompt": "How do I deploy a Stylus contract?"}'

💬 Web Interface

You can also test the AI assistant using a simple graphical interface:

👉 https://stylus-demo.duckdns.org/assistant

No setup or code required — just open the link and start chatting.