The Turnlogue blog
Notes from inside the conversation
Product news, engineering deep dives, and what we’re learning about turning speech into records.

OpinionSeptember 21, 2026 · 5 min read
Meeting bots have a consent problem. You don't have to.
Otter is fighting five consolidated class actions, Fireflies was sued over voiceprints, and Granola's silent transcription landed in court. The industry's default capture model — an AI bot that joins every call — is the root cause. There's a safer architecture, and it's also a better product.
Read the post
ProductSeptember 20, 2026 · 4 min read
Introducing Turnlogue: every conversation, recorded in every language
Meetings, calls and interviews vanish the moment they end. Turnlogue keeps the record — who said what, translated live, summarized by topic. Here's why we built it.
Read the post
EngineeringSeptember 18, 2026 · 6 min read
How live speaker diarization actually works (and why it's hard)
"Who said what" sounds trivial for humans and brutal for machines. A look at streaming diarization, code-switching, and the failure modes we designed Turnlogue around.
Read the post
EngineeringSeptember 16, 2026 · 5 min read
Transcribing a 10-hour recording without falling over
Most transcription services quietly cap out around two or three hours. Here's the checkpointed pipeline that lets Turnlogue take a ten-hour file and return a verified, topic-structured summary.
Read the post