The Git for your AI prompts — version, test, and share prompt chains
Signal what matters · track what ships
Claude 3.5 Sonnet and Gemini 1.5 Pro support
The model picker only has GPT-4 and Claude 3. Need the latest models for meaningful A/B comparisons.
Token count wrong for non-ASCII characters
Japanese and Arabic text shows wildly wrong token counts. The tokenizer seems to be defaulting to ASCII-only counting.
0 comments
No comments yet. Be the first to leave one.