Answer
wavecat runs its models directly on your computer using llama.cpp, so there are no data centers involved.
wavecat can search the web with its web search tools (via Wikipedia, DDGS API), but other than that all processing is done on your computer locally. Again, no personal data ever leaves your computer. You can turn off your WIFI and wavecat will still work. (For more details on local model inference, see here.)
The only tradeoff is that wavecat can be compute-heavy and require a beefy laptop, desktop, or workstation. These PC hardware requirements will lessen over time as models and inference systems become more and more efficient.