Sandbox & Playground

Explore Regolo AI models interactively without writing code.

Regolo Playground

The Regolo Playground is an interactive web interface where you can:

• Test different models - Compare responses across model families
• Experiment with parameters - Adjust temperature, max tokens, and more
• Save conversations - Keep track of your experiments
• Export code - Generate code snippets in your preferred language

Open Playground →

Available Models

In the playground, you can test all available models:

Model Family	Example Models	Best For
Llama	Llama-3.3-70B-Instruct	General chat, reasoning
Vision	Llama-3.2-11B-Vision-Instruct	Image understanding
Embedding	multilingual-e5-large-instruct	Text embeddings
Rerank	bge-reranker-v2-m3	Search result reranking

See the full model catalog for all available models.

Playground Features

System Prompts

Set custom system prompts to control the assistant's behavior:

You are a helpful coding assistant specialized in Python.
Provide concise answers with code examples when relevant.

Parameter Tuning

Parameter	Range	Description
temperature	0.0 - 2.0	Controls randomness (lower = more focused)
max_tokens	1 - 4096+	Maximum response length
top_p	0.0 - 1.0	Nucleus sampling threshold
frequency_penalty	-2.0 - 2.0	Reduces repetition

Conversation History

The playground maintains conversation context, allowing you to:

• Build multi-turn conversations
• Test follow-up questions
• Simulate real-world chat scenarios

API Testing

For direct API testing, check out the interactive API documentation:

API Reference →

The Swagger UI allows you to:

• Browse all available endpoints
• Make authenticated requests
• View request/response schemas
• Download OpenAPI specification

Tips for Experimentation

1. Start simple - Begin with basic prompts before adding complexity
2. Compare models - Same prompt, different models, different results
3. Iterate on prompts - Small changes can significantly improve outputs
4. Save successful prompts - Export configurations for production use
5. Monitor tokens - Watch token usage to optimize costs