Changelog

A public record of what shipped.

SpeechTurn is moving quickly. This page tracks the product changes that matter: the shift from AI question packs to conversation craft support, the public demo workspace, exports, starter rooms, and the quality work behind the brief.

View roadmapOpen demo workspace

How to read this

The changelog is intentionally plain. It is here to show momentum, not to inflate small edits. Major product changes get grouped by month; smaller fixes appear when they change the user experience.

May 2026

Product reset

SpeechTurn becomes conversation prep, not question generation.

We rebuilt the product around conversation rooms, room understanding, and host-ready conversation moves. This is the version where SpeechTurn started to feel like a real preparation artifact.

Rebuilt SpeechTurn around conversation rooms instead of generic question packs
Added room setup for podcasts, panels, interviews, and firesides
Added participant context, reference links, and room understanding
Introduced conversation moves, cadence plans, follow-up ladders, and recovery pivots
Redesigned the brief experience around host control
Added public demo workspace with fictional sample rooms
Added starter-room access flow
Added export options for generated briefs
Improved billing and access states
Switched UI icons to Phosphor
Improved SEO and GEO foundations with sitemap, robots, metadata, and llms.txt
Added broader user feedback and toast states
Reviewed core security posture around Supabase, OpenAI, and Firecrawl usage

April 2026

Early build

The first preparation workflow comes together.

The early product explored how far AI-generated preparation could go before the larger product reset. This period established the foundation: auth, emails, billing, sessions, and the first brief experiments.

Built the first SpeechTurn preparation workflow
Added early session and question generation experiments
Tested initial brief formats and response path structures
Added Resend emails and participant invite flows
Added Supabase authentication foundations
Added Stripe billing foundations
Started internal tests across podcasts, panels, interviews, and game/media examples

Quality pass next

The next work is polish, not a new product direction.

Make export rendering match the dashboard brief from one shared renderer

Clean up first-time onboarding and starter-room language

Improve mobile polish across room and dashboard surfaces

Tighten generated language so questions sound more speakable

Try it

The changelog is useful only if the product keeps getting clearer.

Open the public demo workspace to see the current product shape across different conversation formats.

View demo rooms