Speakr is a personal, self-hosted web application designed for transcribing audio recordings
Python
Prepare v0.10.8-alpha: bump version badge and update the docs home page
Murtaza Nasirecdf9fce1c6f
diff --git a/README.md b/README.mdindex 6bd213b1..3547b368 100644--- a/README.md+++ b/README.md@@ -9,7 +9,7 @@<a href="https://www.gnu.org/licenses/agpl-3.0"><img alt="AGPL v3" src="https://img.shields.io/badge/License-AGPL_v3-blue.svg"></a><a href="https://github.com/murtaza-nasir/speakr/actions/workflows/docker-publish.yml"><img alt="Docker Build" src="https://github.com/murtaza-nasir/speakr/actions/workflows/docker-publish.yml/badge.svg"></a><a href="https://hub.docker.com/r/learnedmachine/speakr"><img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/learnedmachine/speakr"></a>- <a href="https://github.com/murtaza-nasir/speakr/releases/latest"><img alt="Latest Version" src="https://img.shields.io/badge/version-0.10.7--alpha-brightgreen.svg"></a>+ <a href="https://github.com/murtaza-nasir/speakr/releases/latest"><img alt="Latest Version" src="https://img.shields.io/badge/version-0.10.8--alpha-brightgreen.svg"></a></p><p align="center">@@ -212,13 +212,17 @@ Complete documentation is available at **[murtaza-nasir.github.io/speakr](https:- [Troubleshooting](https://murtaza-nasir.github.io/speakr/troubleshooting) - Common issues and solutions- [FAQ](https://murtaza-nasir.github.io/speakr/faq) - Frequently asked questions-## Latest Release (v0.10.7-alpha)+## Latest Release (v0.10.8-alpha)++**A fix for "Invalid model size ''" at startup with whisperx-asr-service (#409).** At startup, a short bundled clip is transcribed to identify the voice embedding model, and no model name was sent with that request, so an ASR service without a default model rejected it. The configured default transcription model is now sent, as for an ordinary upload, and with whisperx-asr-service 0.4.2, `large-v3` is used when `PRELOAD_MODEL` is empty. Incognito mode and bulk reprocessing now use the configured model settings, and a container started from an unchanged example configuration no longer stops on the example admin address. New tests start Speakr and whisperx-asr-service from their documented settings, so problems of this kind are caught before a release. **Full release notes on the [GitHub release page](https://github.com/murtaza-nasir/speakr/releases/tag/v0.10.8-alpha).**++### v0.10.7-alpha (previous release)**"Archived" now means something you choose: an archived recording leaves the main list and nothing is deleted.** Archive a recording from the icon next to inbox and star, or several at once from the selection bar; search still finds archived recordings, and the Archived toggle below the search bar shows only them. Archiving is personal, so archiving a recording shared with you hides it from your list only (#394). The state Speakr previously called "Archived", where audio retention deleted the media and kept the transcript, is now called **Audio removed** and has its own quick filter. You can also put a single recording into that state yourself with **Delete audio, keep transcript** (for a video, **Delete video, keep transcript**). If you used the old Archived toggle to find recordings whose audio retention removed, use the Audio removed filter instead.Several files can now be joined into one recording directly in the upload dialog: choose **One recording**, check the order, and the audio is joined and transcribed once. The recording header groups its actions: inbox, star and archive; folder and tags; a Reprocess menu and Identify Speakers; then share and a ⋯ menu with both deletes. The Identify Speakers dialog is rebuilt for recordings with many detected speakers: speakers sorted by speaking time, short fragments grouped and mergeable, a filter to one speaker's segments, name suggestions from your saved speakers, and a sample button per speaker (#395). AI title instructions are configurable per tag, folder, user and deployment, with the same precedence as the summary prompt (#400). Voice matching is rebuilt: a profile is now a set of samples forming one or more voice variants per person, corrections and removed samples undo themselves completely, profiles are kept separately for each embedding model, short or unlike samples are kept out, and match thresholds calibrate themselves from your data. OpenASR can now return speaker embeddings for voice profiles (#380). Also fixed: SQLAlchemy 2.1 would have broken every PostgreSQL connection on the next image build (#401, fixed by @jjsmackay), the installed desktop app could not be dragged (#402, @jjsmackay), actions behind a proxy that strips the Referer header now report why they fail (#388), and deleting a recording through API v1 now removes its media file. Database columns migrate automatically; no configuration changes are required. **Full release notes on the [GitHub release page](https://github.com/murtaza-nasir/speakr/releases/tag/v0.10.7-alpha).**-### v0.10.6-alpha (previous release)+### v0.10.6-alpha**Background work now runs once per installation instead of once per worker, plus large uploads through restrictive proxies and opt-in email notifications.** Speakr's startup code runs in every process that imports the application, so with the three gunicorn workers the image ships, each of them started its own copy of the background machinery. For the job queue that meant a process booting while another was mid-transcription could un-claim that live job and send the same audio to the ASR service again, once per worker, with the transcript still arriving correctly so nothing looked wrong (#384, reported and fixed by @jagd700). The same pattern affected the webhook dispatcher, the watch-folder monitor and the retention and cleanup schedulers. One process is now elected to own each of these, the rest serve requests and enqueue normally, and raising `--workers` for request capacity is safe. **Full release notes on the [GitHub release page](https://github.com/murtaza-nasir/speakr/releases/tag/v0.10.6-alpha).**diff --git a/VERSION b/VERSIONindex f6064910..0ad9f4d3 100644--- a/VERSION+++ b/VERSION@@ -1 +1 @@-v0.10.7-alpha+v0.10.8-alphadiff --git a/docs/index.md b/docs/index.mdindex 264860ca..ba582869 100644--- a/docs/index.md+++ b/docs/index.md@@ -6,10 +6,10 @@ Speakr is a powerful self-hosted transcription platform that helps you capture,<img src="assets/images/screenshots/main-view.png" alt="Main Interface" style="border-radius: 8px; box-shadow: 0 4px 12px rgba(0,0,0,0.1);"></div>-!!! success "Latest Release: v0.10.7-alpha: archive, joining files at upload, and rebuilt voice matching"- Archiving now hides a recording from the main list without deleting anything, and the state previously called "Archived" (audio removed by retention, transcript kept) is now **Audio removed**, which you can also apply to one recording by hand. Several files can be joined into one recording directly in the upload dialog. Voice profiles are rebuilt from per-recording samples, kept separately per embedding model, with self-calibrating thresholds. The Identify Speakers dialog is rebuilt, and AI title instructions are configurable per tag, folder, user and deployment. Database changes migrate automatically; no configuration changes are required.+!!! success "Latest Release: v0.10.8-alpha: startup model fix for whisperx-asr-service"+ Fixes "Invalid model size ''" at startup with whisperx-asr-service when `PRELOAD_MODEL` is not set (#409): the configured default transcription model is now sent with every request, including the voice embedding check. Incognito mode and bulk reprocessing use the configured model settings, and the admin account is created from the example configuration unchanged. No configuration changes are required.- See the [full release notes](https://github.com/murtaza-nasir/speakr/releases/tag/v0.10.7-alpha) for details.+ See the [full release notes](https://github.com/murtaza-nasir/speakr/releases/tag/v0.10.8-alpha) for details.## Quick Navigation@@ -128,6 +128,14 @@ Learn more about [audio synchronization features](user-guide/transcripts.md#audi## Latest Updates+!!! info "Version 0.10.8-alpha - Startup model fix for whisperx-asr-service"+ No database or configuration changes are required.++ - **"Invalid model size ''" at startup (#409)** - The configured default transcription model is now sent with the voice embedding check, as with an upload.+ - **Model settings** - Incognito mode and bulk reprocessing use the configured model, hotwords and speaker counts; empty model names are never sent.+ - **Example admin address** - A container started from an unchanged example configuration no longer stops on `[email protected]`.+ - **Tests for the documented setup** - Speakr and whisperx-asr-service are tested from their documented settings in CI.+!!! info "Version 0.10.7-alpha - Archive, joining files at upload, voice matching, and configurable titles"Database tables and columns are added automatically; no configuration changes are required.