Use logs to analyze, optimize, and fine-tune AI.
Analyze model usage
Query logs to understand usage, trouble shoot problems, calculate COGS, and evaluate models.
Optimize AI features
Enable caching and batching to reduce costs and latency. Test different prompts and models.
Prepare datasets
Generate training sets for fine-tuning. Use the batch API for evaluations, classification, or embeddings.
Switch between models
Test and try different models without losing data. Reduce platform lock-in by warehousing requests.
Email us at team@usevelvet.com.
