Trainkore Howto
Trainkore is an automated prompt engineering platform that enables model switching, evaluation, and optimization across multiple LLM providers while reducing costs by up to 85%.
View MoreHow to Use Trainkore
Install Trainkore: Import and initialize Trainkore in your project using: import Trainkore from 'trainkore'
Create Instance: Initialize a new Trainkore instance with: const trainkore = new Trainkore()
Configure Chat Prompt: Set up your chat prompt by creating an object with messages and model parameters using trainkore.chatPrompt.create()
Select Model: Choose from available models including OpenAI, Anthropic, Llama2 or custom models in the model parameter
Generate Prompts: Use the auto prompt generation feature to dynamically create prompts for different use cases
Monitor Performance: Access the observability suite to view metrics, debug logs, and analyze input/output performance
Version Control: Use the prompt versioning system to manage and iterate on prompts across your organization
Evaluate Results: Review logs containing input, output, ratings, prompts, and metadata to evaluate performance
Trainkore FAQs
Trainkore is a prompting and RAG platform that offers auto prompt generation, model switching, and evaluation capabilities. It provides higher performance and lower cost than using any single LLM.
View More