Claude upgraded to Claude Sonnet 5.5 for Plus subscribers.

Oct 4, 2026

The "Expected speakers" setting is now on the main recording and import screens, making it easier to get accurate speaker labels when three or more people are talking.

Oct 3, 2026

Fixed an issue where some recordings failed to transcribe.

Oct 3, 2026

Team prompts with attached files now keep working for everyone they're shared with.

Oct 2, 2026

Fewer declined or failed responses in chats without attachments. If the main AI models can't answer, a backup model now responds instead.

Sep 28, 2026

If your internet connection drops during a recording, BastionGPT now more reliably transcribes what was recorded and lets you know part may be missing.

Sep 27, 2026

New solo mode for dictation: set Expected speakers to 1 when you're the only one talking.

Sep 19, 2026

You can now turn off automatic speaker names in transcripts.

Sep 19, 2026

Transcripts now appear right away, and speaker names fill in once they're ready.

Sep 19, 2026

Improved transcription processing time, about two minutes sooner on average.

Sep 19, 2026

Dragging a chat into a folder now moves it reliably.

Sep 19, 2026

Recordings of virtual sessions no longer stop unexpectedly when you connect or switch headphones or microphones.

Sep 19, 2026

Added Taiwanese Mandarin as a transcription language.

Sep 19, 2026

Recording with BastionGPT no longer makes you sound muffled to others on a video call at the same time.

Sep 17, 2026

Your typed message is no longer lost if the page runs into an unexpected error.

Sep 12, 2026

Fewer unexpected sign-outs during brief internet interruptions.

Aug 29, 2026

Clearer message when a document is over the 1,000-page limit or appears to be empty.

Aug 25, 2026

Fast transcription option now available for Plus subscribers.

Aug 22, 2026

Improved notification of low microphone volume during transcription.

Aug 21, 2026

Plus subscribers maximum file uploads increased from 10 to 20.

Aug 18, 2026

Improved microphone selection and recording performance on macOS and iOS

Aug 16, 2026

Improved transcription reliability on unstable internet connections

Aug 15, 2026

Clearer warning when you reach the maximum number of attached files

Aug 5, 2026

Improved the reliability of audio recording when users are on an older device, when Bluetooth audio devices are connected mid-recording or where other meeting apps are running at the same time.

July 24, 2026

Upgraded the intelligence of many models.

July 3, 2026

Upgraded the transcription's speech recognition capability for improved quality.

June 27, 2026

Resolved text output formatting issues when responses contained certain special characters.

June 26, 2026

Addition of GPT 5.5 as a response mode

June 23, 2026

Compliance documents added to settings menu

June 21, 2026

More reliable processing of complex documents, such as graphical charts and handwritten annotations.

June 6, 2026

Plus subscribers can configure AI personality and custom instructions across all chats.

Apr 26, 2026

Added support for WMA and FLAC file uploads.

Apr 25, 2026

Option to manually select the number of speakers for transcription, improving speaker recognition.

Apr 25, 2026

Option to enable time stamps on transcripts.

Apr 25, 2026

Speaker names are automatically detected and added to transcripts.

Apr 25, 2026

Visual improvements of the AI scribe, including an unread status indicator.

Apr 25, 2026

Claude is now manually selectable via response mode selection for Plus subscribers.

Apr 11, 2026

Overall performance improvements. AI personality update.

Mar 8, 2026

Major transcription performance and stability with older devices.

Feb 19, 2026