System requirements¶

Requirements depend on what you use: transcription only, cloud API summary, or offline local summary (Qwen).

Operating system¶

Requirement	Detail
OS	Windows 10 or 11 (64-bit)
Admin	Not required for normal use
Internet	Required for first-run model download and optional cloud summaries

Comparison by scenario¶

Scenario	RAM	Free disk	Network at runtime
Transcription only	8 GB min · 16 GB recommended	~8 GB	No (after setup)
Transcription + API summary	Same as above	Same	Yes during summary
Transcription + local Qwen	≥ 16 GB · 32 GB ideal	~10–12 GB	No (after setup)

RAM and local summary

Before local summary the pipeline unloads the ASR model from RAM. Qwibo checks total physical RAM; below 16 GB the local Qwen engine is not offered.

Transcription only¶

Resource	Detail
RAM	Minimum 8 GB; 16 GB recommended for long files (1 h+).
Disk	~2.5 GB ASR model + app data + job output → ≥ 8 GB free.
CPU	Modern x64; expect ~2× realtime on CPU (1 min audio ≈ 2 min processing).
GPU	Not used in the desktop alpha (CPU-only ASR).

Transcription + cloud summary¶

Resource	Detail
RAM	Same as transcription — LLM runs in the cloud.
Disk	No extra model for summary.
Network	Required when generating summaries.
API key	Configure in app → Summary settings.

Suitable for PCs with 8 GB RAM if you use cloud providers only.

Transcription + local summary (Qwen)¶

Resource	Detail
RAM	≥ 16 GB (checked at setup). 32 GB recommended for long texts.
Disk	Transcription footprint + ~2 GB Qwen GGUF.
CPU	Summary is CPU-bound; long texts may take several minutes (map-reduce).

Bundled dependencies (end user)¶

The installer includes embedded Python, ffmpeg, and the Qwibo backend. You do not need to install these separately.

Self-hosted alternative¶

A Linux mini PC with more RAM/GPU can run qwibo-docker instead. Hardware guidance for Docker is documented in that repository.