updates

FlowPick v1.1.0 Release Notes: Smarter Media Detection

FlowPick v1.1.0 ships — grouped video cards, MSE recording, document sniffing, subtitle detection, and a pile of UI fixes. Here's what changed and why.
FlowPick Team
7 min read
# release # v1.1.0 # update # changelog # mse # subtitles

v1.0.0 was "ship the foundation." v1.1.0 is "fix the things that annoyed us and users." This release is less about flashy new features and more about making detection actually reliable on real-world sites — the ones with dynamic players, preloaded manifests, and 15-second ad segments polluting the stream.

If you're already running FlowPick, update from GitHub Releases. If you're new, start with the v1.0.0 release notes — this post assumes you know what FlowPick is.

More reliable media detection

The biggest internal change. v1.0.0's detection was passive — it watched network requests and hoped for the best. v1.1.0 adds active detection paths for dynamic media pages and preloaded content.

  • Dynamic media pages. Sites that load video via JavaScript after the page renders — previously, FlowPick would miss these because the manifest request happened before the extension was ready. Now we hook into player data.
  • Preloaded content. Browsers preload media metadata for performance; v1.1.0 combines player data with native preloads to catch resources that never hit the network tab in DevTools.
  • Richer metadata. Cards now preserve page title, duration, and cover image whenever they're available — no more generic "media-resource-1" labels.

One video, one card

This was the #1 user complaint. On DASH streams (and some HLS), audio and video tracks arrive as separate resources. v1.0.0 displayed them as unrelated items — you'd see "video_1080p.mp4" and "audio_128k.m4a" next to each other and have to figure out they belonged together.

v1.1.0 groups them:

  • Expandable video cards. Audio and video tracks for the same stream collapse into one card with a expand/collapse toggle.
  • Quality variants inside the group. 1080p, 720p, 480p all live under the same card as selectable variants.
  • Audio track retention. When you pick a quality, the matching audio track is held for the download hand-off — no manual pairing needed.
  • Cover art fallback. If a site provides artwork, the card uses it. If not, we grab the video's first frame for both the card thumbnail and hover preview.

This is the feature that should've shipped in v1.0.0. Lesson learned.

Smoother browsing

Bug fixes for the media list — the stuff that made the extension feel janky.

  • Duplicate stream cards fixed. The "All" tab was showing grouped stream cards twice because of a deduplication bug. Gone.
  • Blank space on long lists fixed. Scrolling long All-tab lists with stream groups would leave large blank gaps — a virtualization issue. Resolved.
  • Image waterfall jumps reduced. Thumbnail loading was causing the image grid to reflow constantly. Now it stabilizes as thumbnails load.

More capture control

New settings and detection paths power users asked for.

  • Hide streaming segments. A toggle to hide individual .ts/.m4s fragments from the list — keeps the UI focused on downloadable resources instead of raw segment noise.
  • MSE recording. Media delivered through the browser's Media Source Extensions pipeline (common on DRM-adjacent and custom players) can now be captured. This was a significant chunk of "FlowPick doesn't see anything on this site" reports.
  • data:image URLs. Inline base64 images now get detected — previously only src attributes pointing to URLs were caught.
  • Document sniffing. Excel, Word, PowerPoint, and other Office files served by the page are now detected. Not the main use case, but useful for corporate intranet pages and course platforms.
  • Subtitle detection. Caption files delivered alongside video (.vtt, .srt references in the manifest) now show up as resources.

Download hand-off reliability

FlowPick's extension detects and organizes; the actual download and merge happen on the FlowPick web downloader page. v1.1.0 improves the data hand-off between the two:

  • Media metadata now transfers safely between extension and downloader — previously, special characters in titles or long manifest URLs could break the hand-off.
  • The extension continues to handle detection and organization; downloading and A/V merging remain on the web side. This separation is intentional — it keeps the extension lightweight while letting the web page handle the heavy FFmpeg work.

What's next

v1.1.0 closes most of the "doesn't work on specific site" gaps. The next release will focus on format conversion (TS→MP4 is there, but MKV/MP3 export is requested often) and performance — merging a 200-segment 4K stream still takes longer than we'd like.

If you hit a site where v1.1.0 still doesn't detect anything, file an issue with the URL and browser version. The changelog page tracks every release.