Terracotta is an easy-to-use platform that streamlines the LLM development workflow. Manage all your fine-tuned models on Terracotta and rapidly iterate improvements through qualitative and quantitative evaluations. It also supports connections with multiple providers, including OpenAI and Cohere. Terracotta facilitates fine-tuning of LLM models by uploading your data and provides secure data storage. Users can categorize data and fine-tune for text generation. Terracotta offers both qualitative and quantitative evaluation features. You can input prompts for multiple models simultaneously and compare the outputs, or utilize our tools to evaluate models across various metrics like accuracy, BLEU, and confusion matrices. Terracotta was co-founded by Beri Kohen and Lucas Pauker, two AI graduate students at Stanford University. Subscribe to our mailing list to stay updated on our latest advancements!