Use Claude Code with Pluely

Connect Claude Code — Anthropic's AI CLI — and every question you ask in the Pluely overlay runs through your own Claude subscription, locally on your machine. No Pluely-hosted models involved.

1. Install and sign in

If you don't have it yet, install Claude Code from Anthropic (for example npm install -g @anthropic-ai/claude-code) and run claude once in your terminal to sign in. If claude answers in your terminal, Pluely can use it.

2. Connect it in Pluely

1. Open the dashboard → Dev space → Connect your CLI.
2. Pluely detects Claude Code automatically (with its version). Click Connect.
3. It appears at the top of the overlay's model picker as Your CLI.

Just installed it? Tap the refresh button beside the "Connect your CLI" heading to re-scan.

3. Pick a model

A Model panel appears right under the connected Claude Code row:

• Default (CLI decides) — uses whatever model your Claude Code is configured for. Always works.
• Sonnet / Opus / Haiku — the latest model of each family, always current (they're Claude Code's own aliases).
• Custom model id — type any exact model id if you need a specific version.

Your choice shows on the connection card and in the overlay picker (e.g. Claude Code · opus).

Images work

Claude Code is vision-capable through Pluely: attach a screenshot or image to your question and Claude reads it. The Capture, Selection, and "Use image" controls all stay enabled.

Thinking models work

Extended-thinking output is handled automatically — you only ever see the final answer, never raw chain-of-thought.

Troubleshooting

• "Your CLI couldn't answer" — run claude -p "hello" in your terminal. If it fails there (expired login, no subscription), fix it there and Pluely follows.
• A custom model id errors — check the id exists for your account; Default always works.

Related

• Connect a local CLI (all options)
• Choosing your AI
• Dev Space overview