v4.1.0Home

Vixen overlay (⌘K)

Press ⌘K (or Ctrl-K) anywhere to open the vixen overlay — a focused inline pane for asking, generating, or fixing.

Surface

┌─ vixen · ask anything ──────────────────┐
│                                                  │
│  > _                                             │
│                                                  │
│  Recent:                                         │
│   ↳ why did nginx fail to start                  │
│   ↳ install postgres and create a dev db         │
│   ↳ explain this command: sed -i 's/foo/bar/g'   │
│                                                  │
│  Tab to accept · Esc to close · ⏎ to run         │
└────────────────────────────────────────┘

What the vixen can do

  • Answer plain questions about the shell, the system, or recent commands.
  • Generate fur commands from natural-language tasks — with explicit confirmation before they run.
  • Explain a previous error and suggest a fix.
  • Write multi-step scripts inline; you can edit before running.

Backend

By default the vixen talks to a local Ollama instance. If Ollama isn't installed, the overlay tells you and links the install hint. A remote backend can be wired up:

bash
fur config set vixen.backend ollama          # default
fur config set vixen.model  llama3.1:8b

# Or a remote endpoint:
fur config set vixen.backend remote
fur config set vixen.remote  https://api.foxora.ai/v1

Confirmation by default

Generated commands always require an explicit ⏎ to run. Auto-execute can be enabled per-session with --vixen-auto, but never in scripts — vixen generation is intentionally interactive.