FaceFrame 1.3.0

What's new in 1.3.0

• Runs smoothly on small PCs. FaceFrame reads your PC (processor, memory, graphics card), picks the best quality it can run without lag, and keeps adjusting while it runs. See Quality → Performance.
• Jump to time for past broadcasts: type hours, minutes and seconds and press Go.
• Faster processing for every PC, with the same picture quality.

Full details: Changelog · Test report

FaceFrame 1.2.0: updates itself

🔄 FaceFrame now updates itself

From this version on you never need to download FaceFrame again. When a new version is published, FaceFrame shows a pop-up at start:

• Update now: downloads and verifies the new version, closes, updates (your settings are kept) and reopens. It takes about 40 seconds.
• Skip: nothing changes, and FaceFrame asks again the next time you open it.

You can also check manually: Diagnostics → Check for updates.

Using 1.1.0? Install this version once with the download link below. Updates after that come through the app.

⬇️ Download

FaceFrame-Setup.exe (102 MB) · Windows 10 (1903+) / 11, 64-bit. It installs everything automatically.

Fixes

• The relaunched FaceFrame.exe could fail with "Failed to load Python DLL" (launcher environment leak).

Full test report: docs/TESTING.md

FaceFrame 1.1.0

⬇️ Download

FaceFrame-Setup.exe (102 MB) · Windows 10 (1903+) / 11, 64-bit

Run it and follow the wizard. Everything else is installed automatically: a private Python runtime, all libraries, the GPU AI runtime on NVIDIA PCs, and the AI models, followed by a self-test. No admin rights needed.

SmartScreen may say "Windows protected your PC" because the installer isn't code-signed. Click More info → Run anyway.

What's new in 1.1.0

• 📡 Twitch link source: full 1080p60 decode for live channels and past broadcasts (?t= start time), low latency, ads skipped, automatic reconnect
• ✨ Best quality: motion-compensated multi-frame fusion + Real-ESRGAN on the GPU (~11 ms per frame)
• 🔊 In-sync stream audio: shows up as "FaceFrame" in the Volume Mixer, with device choice, volume and mute
• ⏯️ Player bar: pause/resume, position / length / remaining time
• 🎨 New interface with an app icon
• 🧰 Windows installer with GPU-aware setup, Repair entry and clean uninstall

Tested

• 100 automated tests, plus a 10-minute live soak on a 1080p60 channel: ~30 fps, 0 rebuffers, stable memory, uninterrupted audio. Long pause and connection drops recover in ~10 s.
• Clean installs (NVIDIA and CPU-only) and uninstall were verified. Full report: docs/TESTING.md

Requirements

• NVIDIA GPU with driver 528+ (RTX 50 series: 570+) for AI enhancement. Without one, Fast mode is used, still at full stream quality.
• About 6 GB of disk with the GPU runtime (about 1 GB without), and internet during setup.
• To send sound into TikTok LIVE Studio: VB-Audio Virtual Cable (free).

Guides: User guide · Troubleshooting

SHA-256 checksums are in SHA256SUMS.txt.