Skip to content
Glade
Esc
↑↓navigate↵open⌘Jpreview
On this page
llms.txt

Claude Code

Run Anthropic's Claude Code in Glade with your own Claude account. What it needs, how to sign in, and what works.

Glade runs your installed Claude Code with the Claude account it’s signed in to.

Requirement Value
Minimum version 2.1.267
Install Claude Code installation
Sign in claude, or Sign in in Settings → Providers

Set up

Install Claude Code

Follow Anthropic’s guide, then check that claude runs in a fresh terminal.

Sign in

Run claude and sign in, or use Sign in next to Claude in Settings → Providers, which runs the same sign-in inside Glade.

Pick a model

Choose a Claude model in the model picker. Effort, fast mode and thinking appear when the model supports them.

What works

  • Steering: a message sent mid-turn can redirect the run.
  • Native forks from any finished reply.
  • Context meter and compaction: /compact and Compact now run Claude’s own compaction, and Claude Code keeps its automatic compaction settings.
  • Approve for me on models and accounts that support Claude’s auto mode.
  • Continue after a usage limit: Glade picks the chat back up when your limit resets.
  • Skills from the shared folder, next to Claude’s own. Invoke them with /.
  • Subagents Claude starts show up in the chat with their own transcripts.

Good to know

  • Your Claude Code settings still apply.
  • Glade keeps Claude Artifacts off, so /design and /slides don’t appear and nothing is published to claude.ai.
  • Deleting a chat in Glade also deletes its Claude session.

Under the hood

Glade drives Claude Code through the Claude Agent SDK. Shared skills reach it through a small local plugin that links the original folders, and Glade’s tools and your MCP servers arrive as MCP servers in each session.