Skip to content

Speakr is a personal, self-hosted web application designed for transcribing audio recordings

Python

Bump version to v0.9.1-alpha

Murtaza Nasir49e3908eaf99
diff --git a/README.md b/README.md
index 003d258e..1ae15200 100644
--- a/README.md
+++ b/README.md
@@ -193,7 +193,11 @@ 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 (v0.9.0-alpha)
+## Latest Release (v0.9.1-alpha)
+
+**A patch release hardening the v0.9.0 upload path.** Fixes uploads failing with an expired CSRF token after long sessions or sleep (#310), Inquire embeddings not being generated when auto-summarization is enabled (#305), and the Account page's API token modals not opening (#308); adds a timeout so stalled uploads fail into the recovery path and a warning before leaving the page mid-upload. **Full release notes: [`release_notes_v0.9.1.md`](release_notes_v0.9.1.md).**
+
+### v0.9.0-alpha highlights (the major feature release this patches)
**The first non-patch release in the v0.8 line.** Three big user-facing themes: capturing audio is now multi-platform and properly documented, the mobile app is a first-class member of the design system, and the upload modal stops feeling like a desktop card pasted onto a phone. **Full release notes: [`release_notes_v0.9.0.md`](release_notes_v0.9.0.md).**
diff --git a/docs/index.md b/docs/index.md
index eb81cc5c..51d66427 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-video.png" alt="Main Interface" style="border-radius: 8px; box-shadow: 0 4px 12px rgba(0,0,0,0.1);">
</div>
-!!! success "Latest Release: v0.9.0-alpha — Multi-platform recording, Stats tab, mobile rebuild, design-system unification"
- The first non-patch release in the v0.8 line. Three big user-facing themes: capturing audio is now multi-platform with a per-OS help guide and a Web Audio mixing path for capturing both sides of a meeting; a new Stats tab shows per-recording metrics; the mobile detail view is a first-class member of the design system. Upload modal redesigned, inquire mode polished, dark-mode select dropdowns finally legible.
+!!! success "Latest Release: v0.9.1-alpha — Upload-path fixes"
+ A patch release hardening the v0.9.0 upload path. Fixes uploads failing with an expired CSRF token after long sessions or sleep, Inquire embeddings not being generated when auto-summarization is enabled, and the Account page's API token modals not opening; adds a timeout so stalled uploads fail into the recovery path and a warning before leaving the page mid-upload.
- See the [full release notes](https://github.com/murtaza-nasir/speakr/blob/master/release_notes_v0.9.0.md) for the complete list. Backwards compatible with v0.8.x; database migrations run automatically.
+ See the [full release notes](https://github.com/murtaza-nasir/speakr/blob/master/release_notes_v0.9.1.md) for the complete list. Backwards compatible with v0.8.x and v0.9.0; database migrations run automatically.
## Quick Navigation
@@ -128,6 +128,16 @@ Learn more about [audio synchronization features](user-guide/transcripts.md#audi
## Latest Updates
+!!! info "Version 0.9.1-alpha - Upload-path fixes"
+ A patch release hardening the v0.9.0 upload path. Backwards compatible with v0.8.x and v0.9.0; database migrations run automatically.
+
+ - **CSRF token expiry on upload (#310)** - The upload path uses `XMLHttpRequest`, which bypassed the fetch-based CSRF refresh, so uploads failed with HTTP 400 once the page token crossed the one-hour limit. It now refreshes the token before sending and retries once on a CSRF rejection.
+ - **Inquire embeddings with auto-summarization (#305)** - Semantic-search chunks were only built in the non-summary path, so with auto-summarization enabled new recordings were never embedded. Summary completion now runs the same chunking step. Pre-existing recordings need a one-time "Re-embed all".
+ - **API token modals (#308)** - An unclosed `<div>` nested the Create Token modal inside the hidden folder modal; the markup is fixed so it opens again.
+ - **Stalled-upload timeout & leave-page warning** - A size-scaled `XMLHttpRequest` timeout routes a stalled upload into the recovery path instead of hanging, and the browser now warns before you leave the page with an upload still in flight.
+
+ See the [full release notes](https://github.com/murtaza-nasir/speakr/blob/master/release_notes_v0.9.1.md) for the complete list.
+
!!! info "Version 0.9.0-alpha - Multi-platform recording, Stats tab, mobile rebuild, design-system unification"
The first non-patch release in the v0.8 line. Three big user-facing themes: capturing audio is now multi-platform, the mobile app is a first-class member of the design system, and the upload modal stops feeling like a desktop card pasted onto a phone. Backwards compatible with v0.8.x; database migrations run automatically.
diff --git a/release_notes_v0.9.1.md b/release_notes_v0.9.1.md
new file mode 100644
index 00000000..f4296245
--- /dev/null
+++ b/release_notes_v0.9.1.md
@@ -0,0 +1,18 @@
+# Release Notes - v0.9.1-alpha
+
+A patch release that hardens the upload path introduced in v0.9.0 and fixes two bugs reported against it. No new features, no configuration changes.
+
+## Bug Fixes
+
+- **Uploads no longer fail with an expired CSRF token.** The file/recording upload path uses `XMLHttpRequest` for progress reporting, which bypassed the fetch-based CSRF refresh. Once the page's token crossed Flask-WTF's one-hour limit (after a long recording, a backgrounded tab, or laptop sleep) the upload returned HTTP 400 and the retry kept failing with the same stale token. The upload path now refreshes the token immediately before sending and retries once on a CSRF rejection, mirroring the existing fetch interceptor. Fixes #310. (Thanks to @fcatuhe, PR #302.)
+- **Inquire embeddings are now generated when auto-summarization is enabled.** Semantic-search chunks were only built in the non-summary completion path, so with auto-summarization on, new recordings completed without ever being embedded and stayed invisible to Inquire mode until an admin ran "Re-embed all". Summary completion now runs the same chunking step. Recordings created before this release still need a one-time re-embed; new uploads embed automatically. Fixes #305. (Thanks to @checkmeck for the diagnosis.)
+- **API token modals open again.** An unclosed `<div>` in the Account page nested the Create Token and token-secret modals inside the hidden folder modal, so clicking Create Token did nothing. Fixes the markup so the token modals render. (Thanks to @fcatuhe, PR #308.)
+
+## Upload Resilience
+
+- **Stalled uploads now fail into the recovery path instead of hanging.** `XMLHttpRequest` had no timeout set, so a dropped-but-not-closed socket fired neither `ontimeout` nor `onerror` and the upload hung forever. A size-scaled timeout (a ten-minute floor plus one minute per 10 MB) now routes a stall through the same failure and recovery handling as any other upload error. (Thanks to @jjsmackay, PR #306.)
+- **A warning now appears before leaving the page during an in-flight upload.** Closing or reloading the tab mid-upload triggers the browser's "leave site?" prompt when any upload is still queued, ready, or uploading. (Thanks to @jjsmackay, PR #307.)
+
+## Compatibility
+
+Backwards compatible with v0.8.x and v0.9.0. Database migrations run automatically on startup. Upgrade with the usual `docker compose pull && docker compose up -d`.