> ## Documentation Index
> Fetch the complete documentation index at: https://docs.stagewhisper.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Changelog

> Everything we shipped, when we shipped it.

StageWhisper works during your calls, not after. No bots join your meetings.
Here is every update since day one.

***

## April 27, 2026

**Now in your menu bar**\
On Mac, StageWhisper runs from the system tray. The control panel stays out of the way until you need it, and the tray menu gives you quick links to the dashboard, settings, and the panel itself.

**Quieter app launch on Mac**\
We stopped requesting macOS permissions the moment you open the app. The OS prompt only fires when you do something that actually needs it, like clicking Listen for Screen Recording.

**A new Permissions tab in Settings**\
You can see and manage every macOS permission StageWhisper uses in one place. Granted, denied, or never asked, all visible. One click takes you to System Settings if you want to change something.

**Heads-up before the Screen Recording prompt**\
The first-run tour now warns you about the macOS Screen Recording dialog right before you click Listen, so you know what to expect.

## April 22, 2026

**Safer sign-in across web and desktop**\
We tightened the auth flow across the web app and desktop app. Sign-in, device linking, and session handling are less likely to leave you stuck in a broken in-between state.

**More reliable BYO assistant pairing**\
We also tightened the OpenClaw pairing path with better heartbeat checks, tighter limits, and a cleaner encrypted handoff. Connected assistants should hold up better during live sessions.

## April 20, 2026

**Feature flags for staged rollouts**\
The desktop app can now receive feature flags, so new capabilities can roll out gradually instead of appearing half-finished for everyone at once.

**Cleaner desktop behavior on Mac**\
We improved panel activation, key handling, and release packaging. That makes the floating controls feel more predictable and helps the app ship with the pieces it needs, including FFmpeg in the desktop release flow.

## April 19, 2026

**Encrypted BYO reasoning path**\
StageWhisper now encrypts BYO reasoning payloads before they leave your device for OpenClaw. That gives local-first users a much tighter privacy boundary during live coaching.

**Smoother local mode and OpenClaw setup**\
We cleaned up the plugin path, tightened compatibility, and made local mode status clearer in the desktop app. Founders Edition users running OpenClaw should now find the connection easier to trust.

## April 16, 2026

**Screen reading joins live context**\
StageWhisper can now read the text on your screen during an active session, only when you switch it on, and use it to improve live suggestions. It stays tied to the session, with clearer controls and status handling in the desktop app.

**Better on-screen context retrieval**\
We added local storage for compacted recordings, window context, and accessibility snapshots so StageWhisper can pull in what is actually on screen while the conversation is still happening.

## April 14, 2026

**Add files to your playbooks**\
You can now attach PDFs, DOCX files, notes, and CSVs to a playbook. StageWhisper keeps those files on your device and pulls in the most useful parts during the call, so you get live help with no bot joining the meeting.

**Connect your calendar for meeting prep**\
You can now link Apple Calendar or Google Calendar and choose how early prep shows up. Your calendar data stays on your device, so StageWhisper can get you ready before the call and stay useful while the call is happening.

**Get a clearer plugin update nudge**\
If the StageWhisper plugin is missing or out of date, the app now shows the exact command to install or update it. That makes it easier to keep live, during-call actions working without extra setup guesswork.

**Make checkout more reliable**\
We fixed a Stripe billing issue so starting or managing a paid plan should be smoother. That helps you get StageWhisper ready before your next call instead of fixing billing after the fact.

## April 9, 2026

**Founders Edition launch**

We launched the Founders Edition: pay \$399 once for lifetime usage, lifetime security upgrades, and one year of free feature upgrades. Early supporters keep everything.

Persona-specific landing pages now show how StageWhisper fits sales, consulting, recruiting, and research workflows.

***

## April 7, 2026

**Local-first transcription with Parakeet ASR**

StageWhisper can now transcribe calls on your Mac. Audio stays on your machine. We use NVIDIA's Parakeet TDT model through ONNX Runtime, so nothing leaves your laptop.

Added a waitlist page for Windows and Linux users.

***

## April 6, 2026

**Founders Edition pricing tier**

Released the Founders Edition: pair StageWhisper with your own OpenClaw instance for live suggestions, custom personas, and playbooks. Published docs at stagewhisper.io/docs.

***

## April 4, 2026

**Durable task execution with Hatchet**

Moved background AI reasoning to Hatchet's durable execution engine. Tasks survive server restarts, retry on failure, and run with proper deadlines.

Tightened rate limiting and security around user tokens.

***

## April 3, 2026

**Hatchet integration and stale job cleanup**

Replaced the old in-memory sweeper with Hatchet-managed workflows. Stuck background jobs now time out cleanly instead of hanging.

Added retry logic for callback event delivery.

***

## April 2, 2026

**Bring your own AI assistant**

You can connect your own AI reasoning service to StageWhisper. If your service goes down mid-call, StageWhisper falls back to hosted reasoning so coaching never stops.

Health checks detect disconnects fast and reconnect when the service comes back.

***

## April 1, 2026

**BYO assistant reliability and session stability**

Fixed a race condition where reconnecting a BYO reasoning service mid-call could hang the session. Switched to SHA-256 hashing for session tokens.

Addressed 15 review findings from the BYO assistant pull request.

***

## March 30, 2026

**OpenClaw relay hardening**

Added a confirmation dialog before disconnecting your AI assistant. Relay message delivery now validates payloads, uses stable idempotency keys, and retries once on failure.

***

## March 27, 2026

**Real-time pub/sub for AI assistant bridge**

Replaced polling with PostgreSQL pub/sub for the AI assistant bridge. Messages from OpenClaw now arrive instantly instead of on a two-second delay.

Published the OpenClaw-StageWhisper channel as a standalone package.

***

## March 26, 2026

**OpenClaw assistant dispatch**

Connected OpenClaw as a live reasoning provider. StageWhisper sends conversation context to your OpenClaw instance during the call and shows responses on screen in real time.

***

## March 24, 2026

**Playbook framework**

Built the playbook system. Define step-by-step frameworks (MEDDPICC, BANT, custom checklists) and StageWhisper tracks your progress through each step during the call.

Added playbook templates for common sales and consulting workflows.

***

## March 22, 2026

**Resumable sessions**

If your app closes mid-call, you can pick up where you left off. Sessions resume with the full transcript and coaching state intact.

Fixed race conditions around session segment tracking and async desktop events.

***

## March 20, 2026

**Subscription management and pricing page**

Added a pricing page showing what each plan includes. Users can manage their subscription, change plans, and view invoices from the app.

***

## March 18, 2026

**About page and SEO**

Launched the About page. Applied SEO metadata across all landing pages.

***

## March 15, 2026

**Privacy controls and desktop consent**

Added privacy consent flows to the desktop app. You choose what audio gets captured and what gets shared before anything leaves your machine.

***

## March 14, 2026

**Polished landing page**

Redesigned the landing page with feature reveal animations, mobile-responsive navigation, and gated downloads that verify you are human before handing over the installer.

***

## March 12, 2026

**Personality voice setup**

You can now set up your coaching personality by speaking to it. Tell StageWhisper how you want feedback and it adjusts its tone and style.

Added toggle controls for individual insight signals.

***

## March 11, 2026

**Live STT-LLM coaching pipeline**

Shipped the core real-time pipeline: speech-to-text feeds into the language model during your call. Suggestions show up on screen while you talk. No waiting until the call ends.

Added Silero VAD for voice activity detection so the system only processes speech, not background noise.

***

## March 10, 2026

**Vocal tone analysis**

StageWhisper reads vocal tone in real time. It tracks eight dimensions (hesitation, confidence, tension, momentum, and more) and shows you how the conversation feels as it happens.

***

## March 7, 2026

**Visual overhaul and session panel**

Replaced the landing page gradient with an animated blinds shader. Updated brand copy.

Widened the session panel, simplified the tone legend, and stopped sending raw mic audio over the websocket. Only transcripts travel to the server now.

***

## March 5, 2026

**Custom personas**

You can create custom coaching personas. Each persona has its own personality, expertise, and coaching style. Switch personas between calls.

***

## March 3, 2026

**Insight signal pipeline**

Built the signal extraction system. StageWhisper pulls structured coaching signals from the live transcript: commitments, objections, questions, action items, and more. All in real time.

***

## March 1, 2026

**Landing page and local dev server**

First public version of the StageWhisper landing page. Set up the local development server.

***

## February 24, 2026

**Multi-session support**

Users can run multiple coaching sessions and switch between them. Session history is preserved and searchable.

***

## February 21, 2026

**Session recorder and desktop branding**

Added the session recorder: StageWhisper captures the full call and makes a searchable transcript when you finish. Audio stays local.

Set up brand identity with app icons, bundle identifiers, and the first DMG installer for macOS.

***

## February 20, 2026

**Stripe billing and entitlements**

Integrated Stripe for subscription billing. Added webhook handling, entitlement checks, and usage caps.

Users can manage settings from a dedicated window.

***

## February 18, 2026

**Live transcription with Gemini**

Connected Google Gemini as a live transcription and reasoning provider. Transcriptions stream in real time during calls.

***

## February 16, 2026

**User authentication and backend API**

Built the auth system: email/password signup, login, JWT tokens, and account management. Connected the desktop app to the cloud backend.

***

## February 11, 2026

**Keyboard shortcuts and session state**

Added global keyboard shortcuts so you can control StageWhisper without switching windows during a call. Improved session state tracking.

***

## February 9, 2026

**Gemini live provider**

Added Google Gemini as a live reasoning provider with input/output transcription tracking.

***

## February 7, 2026

**Backend API foundation**

Set up the FastAPI backend with database models, migrations, and the core API routes the desktop app talks to.

***

## February 5, 2026

**Transcript panel and suggestion engine**

Built the transcript panel that shows live transcription and coaching suggestions side by side. Added global shortcuts for toggling the panel during calls.

***

## February 3, 2026

**Desktop app foundation**

First working version of the desktop app. Screen sharing privacy controls, live transcript panel, and a clean overlay UI that sits beside your call window.

***

## January 27, 2026

**Development environment setup**

Cleaned up the dev environment and terminal config for consistent builds.

***

## January 25, 2026

**Project foundation merge**

Merged the foundation branch: Tauri desktop shell, React frontend, and the project structure everything builds on.

***

## January 23, 2026

**Tauri desktop app bootstrap**

Initialized the Tauri app with React, Vite, and macOS private API support for system audio capture. This is the shell that became StageWhisper.

***

## January 6, 2026

**Skill system and constitution**

Built the skill installer and validation framework. Defined the product constitution, the rules StageWhisper follows when coaching.

***

## January 5, 2026

**Initial commit**

Created the repository. Day zero.
