Daedalus — powered by the Hermes Deep Mind engine

Thin client for Daedalus (Hermes Deep Mind engine): streaming chat sidebar, native inline diffs for every agent edit (accept/reject via the changeset protocol), active-file context, auto-routing across your local models.

Setup

1. Start the agent (from your project directory):

   pip install daedalus-ai
   daedalus ws        # agent server on ws://127.0.0.1:8765
   

2. Install the extension:

   cd vscode-extension
   npm install
   npx @vscode/vsce package        # produces daedalus-vscode-0.1.0.vsix
   code --install-extension daedalus-vscode-0.1.0.vsix
   

   (Dev mode instead: open this folder in VS Code and press F5.)

3. Click the Daedalus icon in the activity bar and chat.

Features

• Chat sidebar — streamed tokens, shows which provider each answer routed to
• Inline diffs — when Daedalus edits files, a notification per file offers Diff (VS Code's native diff of original vs current), Accept, Reject (reject restores the original content)
• Daedalus: Ask About Current File — sends your question with @<file> attached (contents expanded server-side)
• Daedalus: Cancel Current Run — stops a generation mid-stream

Settings

• hermes.host — WS URL (default ws://127.0.0.1:8765)
• hermes.token — only needed when the server sets HERMES_WS_TOKEN (hermes web prints its token; hermes ws runs open on localhost)