From 2ebccdbc37b3c073d36368238195af0c885faaeb Mon Sep 17 00:00:00 2001 From: ComputelessComputer <63365510+ComputelessComputer@users.noreply.github.com> Date: Fri, 29 May 2026 00:54:57 +0900 Subject: [PATCH] chore(changelog): add 1.0.32 release notes Document user-facing desktop fixes for titles, chat context, audio retention, imports, transcription, and timeline tabs. --- packages/changelog/content/1.0.32.md | 29 ++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) create mode 100644 packages/changelog/content/1.0.32.md diff --git a/packages/changelog/content/1.0.32.md b/packages/changelog/content/1.0.32.md new file mode 100644 index 0000000000..4631fca04f --- /dev/null +++ b/packages/changelog/content/1.0.32.md @@ -0,0 +1,29 @@ +--- +date: "2026-05-29" +summary: "Manual titles are preserved, chat context and storage settings are clearer, and imports, local transcription, and timeline tabs are more reliable." +--- + +## Notes + +- Manually edited note titles now stay intact when title generation or note enhancement finishes in the background +- Notes and calendar events opened from the top timeline now open in tabs, selecting an existing tab when possible instead of replacing the current note + +## Chat + +- The chat context picker now shows recent notes before you search and keeps results aligned with the latest saved note data +- Searching from the chat context picker now shows progress while matching notes are loading + +## Audio & Storage + +- Recording retention now includes a Forever option for keeping audio until you delete it manually +- Fresh installs now keep saved recordings by default, while existing "don't save" choices continue to be honored + +## Import + +- Legacy Hyprnote imports now copy the database before scanning or importing, making imports more reliable while the original database is still changing +- Legacy import scans now use faster counts, and long-running scans or imports now show timeout errors instead of appearing to hang indefinitely + +## Transcription + +- Local Soniqo Parakeet now uses batch transcription for supported non-English languages when live transcription is not the right fit +- Saved audio processed after Soniqo live sessions now uses the batch model for more reliable file transcription