Skip to content

Releases: thedevmark/Clipline

v0.2.6

Choose a tag to compare

@github-actions github-actions released this 04 Sep 06:35

Full Changelog: v0.2.5...v0.2.6

v0.2.5

Choose a tag to compare

@github-actions github-actions released this 03 Jun 04:45

Fixed

  • Twitch VOD/clip download — double-clicking a VOD failed with "[Errno 2] No such file or directory", and downloading two at once could crash. Each download now uses its own isolated folder, drops the --no-part flag that broke Twitch's fragmented downloads, and only one runs at a time.
  • VOD/clip library too small — the list now fills the panel's height instead of a cramped strip.
  • Welcome runtime status — dropped the stale "faster-whisper (pip)" row; now shows the real optional engines, captions (whisper.cpp) and speaker separation (sherpa-onnx), with their actual status. Both are 1-click, no-pip setups in the Shorts stage.

Install: download clipline.exe below and run it.

v0.2.4

Choose a tag to compare

@github-actions github-actions released this 03 Jun 03:41

Polish on the v0.2.3 captions.

Changed

  • Karaoke caption colour — words not yet spoken now show as a dimmed version of each speaker's own colour (brightening to the full colour as each word is spoken), instead of a fixed red. Each speaker reads as one coherent colour.

Install: download clipline.exe below and run it.

v0.2.3

Choose a tag to compare

@github-actions github-actions released this 03 Jun 03:27

Captions you can actually ship: burned into the video, split by speaker, and placed where you want them.

Added

  • Caption burn-in — captions are now rendered into the video on export. The editor's burn-in toggle finally does what it says, and it applies to every clip and the longform build (per-clip, time-shifted so timing stays right).
  • Automatic speaker separation — a one-time ~50 MB download (sherpa-onnx, fully offline, no Python/pip — same one-click model approach as captions) splits the transcript into speakers. Pick Auto / 2 / 3 / 4 speakers; pinning the count is more accurate when you know your cast.
  • Per-speaker colour + on-screen position — give each speaker their own caption colour and drag their captions to a spot on a clip preview; any single caption line can be nudged to its own position. Positions are resolution-independent.

Notes

  • Speaker separation and captions run locally; nothing is uploaded. The diarization models are commercial-use-safe (Apache-2.0 / MIT) — see THIRD_PARTY_NOTICES.txt.

Install: download clipline.exe below and run it. No Python, no pip, no terminal.

v0.2.2

Choose a tag to compare

@github-actions github-actions released this 02 Jun 22:48

Ships the loopback Twitch sign-in to a release and closes two welcome-screen rough edges.

Added

  • "Sign in with Twitch" (loopback) is now the default connect flow in a released build — click Authorize in the browser and it bounces straight back, no code to type. Device-code linking remains as the automatic fallback when the loopback port is busy. (Previously shipped builds predated this and always used device-code.)

Changed

  • Help → About now shows the app version and current capabilities instead of the stale "Phase 2 native build" line.

Fixed

  • Welcome runtime-status "Re-check" now shows a timestamped confirmation ("Checked HH:MM:SS — all required tools found" / "missing: …") so it no longer reads as a dead button when nothing needs installing. Install buttons still appear only for tools that are actually missing.

Install: download clipline.exe below and run it. No Python, no pip, no terminal.

v0.2.1

Choose a tag to compare

@github-actions github-actions released this 01 Jun 21:37

The auth-first ingest batch: connect Twitch, pull your clips, caption them — with zero terminal/pip.

Added

  • Twitch login (OAuth Device Code Flow) with a bundled public client ID — connect out of the box, no setup.
  • Twitch VOD/clip browser in Ingest — double-click to download into the editor.
  • URL ingest — paste any Twitch/YouTube/TikTok URL; yt-dlp fetches with progress.
  • Inbox inspector + waveform timeline with draggable in/out trim handles.
  • One-click captions (whisper.cpp) — first use downloads a ~75 MB speech engine (no Python/pip/terminal); word-level karaoke. Caption editor: edit text, toggle words, speaker colour, ASS/SRT export.
  • First-run guided tour over the timeline / inspector / render button.

Changed

  • Full-bleed app/taskbar icon (was a blurry upscaled favicon); vector HiDPI welcome icon; Windows AppUserModelID set.
  • Welcome runtime-status card scrolls and offers per-tool Install (winget) + Re-check.

Fixed

  • Desktop Twitch auth — device-code flow replaces the dead webview/localhost-rejection path.
  • "Open Local Video…" no longer force-navigates to Ingest on cancel.
  • --selftest exits with the real code instead of crashing on Qt teardown.

Downloaded EXE verified: 64.5 MB, --selftest and --selftest-deps green on a clean run.

v0.2.0-phase3

Choose a tag to compare

@github-actions github-actions released this 24 May 19:31

v0.2.0-phase0

Choose a tag to compare

@github-actions github-actions released this 24 May 19:12

Full Changelog: v0.1.3...v0.2.0-phase0

v0.2.0

Choose a tag to compare

@github-actions github-actions released this 24 May 19:45

Full Changelog: v0.1.3...v0.2.0

v0.1.3

Choose a tag to compare

@github-actions github-actions released this 24 May 18:40

Full Changelog: v0.1.2...v0.1.3