Not Diamond - All-In-One AI Platform - Chat With Any LLM for FREE! (LLM Routing)

---

---

AI Summary

Summary of Not Diamond AI Platform Video

• Introduction to Not Diamond:
  • Not Diamond is a free AI model router.
  • Automatically selects the best language model (LM) for each query.
  • Aims to maximize output quality while reducing cost and latency.
  • Offers state-of-the-art performance by dynamically choosing the most suitable model.
  • Allows for personalized routing based on real-time user feedback.
  • Users can train custom routers, control LM requests client-side, and integrate with Python and TI script via REST API.
• Comparison with OpenAI Models:
  • OpenAI’s GPT-3.5 is powerful but expensive.
  • Not Diamond uses a repository to route to GPT-3.5 only when it makes a significant difference.
• Getting Started with Not Diamond:
  • Users can sign up with an email, Google, or GitHub account.
  • Main chat dashboard is where users interact with the router.
  • Users can toggle between different LMs based on the task at hand.
  • Models include GPT-4 Omni, CLIP 3.5 Sonic, Llama, and Perplexity.
  • The platform supports image capabilities and has an “Arena Battle Mode” for model comparison.
• Arena Battle Mode:
  • Allows comparison of outputs from two LMs side by side.
  • Helps users choose the best model for specific tasks.
  • Example given of generating an iOS calculator app.
• Custom System Prompts:
  • Users can create detailed instructions for models to follow.
  • Prompts can be named and saved for specific tasks.
  • Example given of creating a responsive finance tracking app.
• Local Installation and API Usage:
  • Not Diamond can be installed locally.
  • Requires Python 3.10 or above.
  • Users can create an API and link model API keys for routing.
  • Custom router training is available to optimize model selection.
• Conclusion:
  • Not Diamond intelligently selects the best LM for each query.
  • Delivers high-quality results across various domains and tasks.
  • Recommended for those looking to optimize performance, cost, and latency.

Note: No specific CLI commands, website URLs, or detailed instructions were provided in the text for extraction.