Frequently Asked Questions

Common installation, runtime, and support questions for Whiskers AI.

General

What is Whiskers?

Whiskers is a runtime layer for local AI models. It adds persistence, continuity, commands, logging, and summarization around a local llama.cpp model server.

Does Whiskers include an AI model?

No. You provide your own GGUF model and llama.cpp server.

Does Whiskers send data to the cloud?

No. Whiskers is designed for local-first operation.

What operating systems are supported?

Linux, macOS, and Windows through WSL2 are supported. On macOS, native llama.cpp with Metal is recommended.

Runtime

Does Whiskers survive restarts?

Yes. Session continuity is stored in logs on disk, not inside the browser tab.

What does the summarizer model do?

The summarizer model processes older session content into compact summaries for long-term continuity and recovery.

Can multiple browser windows share the same session?

Yes, as long as they connect to the same Whiskers runtime instance.

Will Whiskers fix hallucinations?

No. Whiskers improves continuity and runtime structure, but the model still generates the text.

Does Whiskers give the model infinite memory?

No. Whiskers uses persistent logs, bounded context, and optional summaries. It improves continuity, but local hardware and model context limits still matter.

Can Whiskers run on an 8GB Mac?

Yes, testing showed Whiskers can run on an 8GB Apple Silicon Mac using native llama.cpp with Metal. Very long or dense sessions may still hit memory limits depending on model size and context settings.

Support Boundaries

What counts as a Whiskers issue?

What does not count as a Whiskers issue?

Licensing

Is Whiskers open source?

Whiskers is source-visible commercial software. Official releases and roadmap direction remain controlled by Chippie Bear Industries, LLC.

Will there be a trial version?

Planned releases currently include a limited evaluation period and optional permanent unlock licensing.