Clippy AI Platform - Test & Embed Widget

⚙️ AI Provider & Model

💡 Universal Model Picker
Select any AI provider supported by LiteLLM (100+ providers!)

🌐 AI Provider

Groq offre accesso gratuito a modelli veloci come Llama, Mixtral e GPT OSS

🤖 AI Model

Select the model from the chosen provider, or type any model code supported by the provider in the field above (it overrides the list).

🔑 API Key

🔒 The API key is encrypted and saved on the server. It is never stored in the browser.

🎨 Interface — Mode & Appearance

🎨 Interface

🎨 Accent color 🌙 Dark mode

👤 Character

🚦 Behavior & Limits

🔍 Enable web search

Only for providers that support it natively (OpenAI, Anthropic, Gemini, xAI, Perplexity). Ignored otherwise.

🚦 Max messages per conversation (0 = unlimited)

✍️ Response length — max output tokens

How long each reply can be. Too low truncates answers mid-sentence; 1200 suits assistants that list detailed results. (50–4000)

🎛️ Temperature — creativity (0 = focused, 2 = wild)

Lower = more precise and repeatable; higher = more varied. 0.7–0.8 is a good default for support bots.

📝 System Prompt

Customize the assistant's behavior:

👋 Welcome Message

Message shown in the balloon when the widget loads:

💡 This message appears in the balloon when the widget first loads. Leave empty for default greeting.

📚 Knowledge Base (RAG)

💡 What is RAG?
Retrieval Augmented Generation allows the AI to respond using information from your documents.

Example: Company FAQs, product manuals, technical documentation

🌐 Fetch from URL:

📂 Upload document: ✏️ Or paste text:

🧪 Tests, Animations & Debug

🎮 Quick Tests

🎭 Animations

🔧 Debug

💾 Save & Generate Embed

1. Save configuration:

Configuration Name:

✅ Configuration saved! ID:

2. Copy this code into your site:

Loading...

👁️ Preview Widget