Changelog

TL;DR: Recent shipping. Newest first. Every entry is dated; subscribe by RSS-watching this page or check back monthly. For older history or release notes about a specific date, contact support.

May 2026

• Production readiness milestones. Internal hardening across rate limiting, session integrity, and concurrency safety so live transcription stays responsive under sustained load.
• Larger-segment support. Removed the 150-token cap that limited segment length on long-form recordings — long monologues now render as single segments instead of being chopped up.
• Quieter session cleanup. Suppressed a noisy warning that fired when empty cleanup beacons arrived from closed tabs; no user-visible behaviour change, cleaner telemetry.
• Mobile sharing. End-to-end tests now cover the mobile sharing journey, locking in a flow that had previously regressed silently on small screens.

April 2026

• Async file transcription is live. Upload an audio or video file — or paste a public URL — and we transcribe it in the background. Close the tab and come back; the session appears in your sidebar when it's done. See Async Transcription for limits and formats.
• AI title, summary, and TOC now follow your interface language. Switch lognow to your locale and AI-generated metadata is produced in that language instead of always English.
• Locale preference sync. Changing your interface language is now persisted across devices and reflected on the next sign-in without manual re-selection.
• Docs surface in marketing. Direct link to documentation from the landing header and signed-in account menu; new async-transcribe doc included from day one.
• Soft-delete for profiles. Account deletion now soft-deletes the profile to prevent signup-credit farming via repeated delete-and-resignup cycles.
• Recording-stopped marker stays at the end. Fixed a layout edge case where the "recording stopped" indicator drifted above the final transcript segment.
• SEO foundation shipped. Site-wide JSON-LD @graph, hreflang clusters for 13 locales, an explicit AI-crawler allow-list, dynamic OG images, and a generated llms-full.txt for LLM ingestion.

Older

• For earlier history, see the git log or contact support.