Choose Faster or More Precise
Pick the Faster (25-language) or Slower, More Precise (100-language) local speech model.
Stellar Recorder is not yet offered as a direct public installer download on this website. Request access via contact to get started on your desktop computer.
Pick the Faster (25-language) or Slower, More Precise (100-language) local speech model.
Selects a supported GPU first (no silent CPU switches if GPU fails; manual CPU option available).
Choose an existing local folder (no network shares) to unlock New Recording or Audio Import.
AI notes stay off until you save and enable an OpenAI-compatible text endpoint and model.
| Component | Requirement & Behavior |
|---|---|
| Operating system | 64-bit desktop operating system. |
| Runtime | Microsoft Visual C++ x64 runtime. If missing, setup halts prior to model activation; the app does not install it silently for you. |
| GPU | Optional. Automatic device selection uses a supported GPU first (including migrated older profiles without a CPU lock). |
| CPU | Used only when no supported GPU is detected or when you explicitly choose CPU mode. |
| Disk space for models | Speech models download explicitly during setup; exact compressed/unpacked download size figures are not yet published here. |
| Project folder | Existing local volume directory chosen by you (SMB/UNC network shares are rejected). |
Stellar Insights has not published a third-party security audit, a code-signing certificate or public release checksums for Stellar Recorder on this website.
When verified public release builds ship here, this section will publish exact version metadata, file sizes and cryptographic SHA-256 checksums for every installer.