HELP WITH LITEREC
LiteRec Support Small Files. Clear Recordings.
Contact: vincent@nordkapp.cn
Please include your LiteRec version, macOS version, and a description of the issue. Avoid attaching private recordings unless they are needed to demonstrate the issue.
Getting started
1. Open LiteRec and allow macOS screen recording access. Quit and reopen the app if macOS requests it. 2. Choose Full display, Selected area, or Window, then select the screen, drag an area, or click a highlighted window. Press Esc or Cancel to leave the selector. 3. Choose Compact, Balanced, or Smooth. Turn System audio on to capture sounds playing on your Mac. 4. Choose your save folder with Change, then click Start recording. 5. Stop and save from the recording timer in the menu bar. Play the MP4 or reveal it in Finder.
Smaller files
Use HEVC and Compact, capture only the area you need, and turn off system audio if it is unnecessary. The displayed estimate is based on target bit rates. Actual size varies; lower settings reduce detail and motion smoothness. For small text, select a smaller area or use a higher quality preset.
Screen recording permission
Open System Settings > Privacy & Security and find Screen Recording or Screen & System Audio Recording, depending on your macOS version. Allow LiteRec and restart it if prompted. If an older copy remains in the list, remove the old entry and add the copy shown by LiteRec's Show this app button.
No system audio
Enable System audio before starting. Play audio in another app and check its volume. Audio may include apps outside the selected area. Protected content may not be captured. This version does not record the microphone.
Purchase verification
Get LiteRec from the Mac App Store using the Apple Account that purchased it or has eligible shared access. If verification is unavailable, connect to the internet and click Verify purchase, or download LiteRec again from your App Store purchases. An Apple authentication prompt may appear. You do not need a separate LiteRec account or a license key.
Offline recording
StoreKit can use an available signed purchase record. A temporary connection failure will not interrupt an already verified session or an active recording. If no valid purchase record is available after launching the app, connect to Apple to verify it first.
Languages
Use the Language menu at the bottom of the recorder to choose English, Simplified Chinese, Traditional Chinese, or the system default. macOS permission dialogs and Finder use your macOS language settings.
Compatibility and playback
Requires macOS 13 or later. MP4 recordings use HEVC or H.264 with optional AAC audio. Choose H.264 before recording for players that do not support HEVC. Audio-only recording saves M4A files. Local transcription uses downloaded models; Base and speaker labeling require macOS 13+, while the included Small and Large presets require macOS 15+. Model inference has been tested on Apple Silicon; performance on older hardware varies.
Purchases and refunds
The Mac App Store handles purchases and refunds. To request a refund, use https://reportaproblem.apple.com/ . Refund decisions are made by Apple.
Transcription
Open Transcribe or press Command-T. Import a local video/audio file, or use Transcribe beside a completed recording. Choose On this Mac, download a speech model and the optional speaker-labeling model, then start. Each file may be up to five hours long; long recordings require more memory and processing time. Supported formats depend on the audio/video decoders supplied by macOS.
Whisper is multilingual, but no model recognizes every language equally well. Choose the source language when automatic detection is unreliable. Base and Small use the standard multilingual vocabulary; the Large v3 preset also includes Cantonese. Model size is shown before downloading. Files come from Hugging Face and are verified; no audio is uploaded during local transcription. Imported model folders must use WhisperKit Core ML format and include tokenizer files; GGUF and MLX packages cannot be imported.
Speaker labels are anonymous. Rename a speaker to update the whole transcript, reassign individual paragraphs, or merge duplicate labels. Click a timestamp to hear that paragraph. Edit text directly and export TXT, Markdown, SRT, VTT or a JSON backup. Automatic labels can be wrong for short, noisy or overlapping speech. The original media must remain available for playback; Locate source restores a moved file. Text remains editable even if the source is missing.
Cloud transcription
Cloud setup contains API presets for OpenAI, Groq and SiliconFlow, a compatible API option, and setup guides for additional providers. Create your own key, enable the provider's billing if required, and save the key in Keychain. ChatGPT subscriptions do not include OpenAI API usage. A cloud job sends only extracted audio after your confirmation. The current upload limit is 24 MB after audio extraction; split larger files or use local mode. OpenAI's diarization preset returns speaker labels; Groq returns timed text; the SiliconFlow preset returns full text without per-segment timestamps. Native Alibaba, Volcengine and Tencent adapters are not included in this release.