Fixed
- The Ask tab in saved meetings could cut off long answers mid-sentence without any indication, because it inherited a 512-token completion budget meant for short structured completions. Ask now gets an 8,192-token budget, and a truncated answer shows a length-limit warning with a Continue action that picks up where it left off. (#727, #728 by @gthay)
Docs
- README's "Pick any LLM" feature bullet now mentions Requesty alongside OpenRouter as a supported cloud provider. (#725 by @Thibaultjaigu)