How does Vesper Code actually work?

Vesper Code runs multiple agents, each tuned for a specific task.

The main agent ("Vesper") reads your prompt, gathers context, and spawns subagents. Publicly, the CLI has one normal working mode and a /plan toggle for planning-only sessions.

• Normal - reads code, edits files, runs commands, and reviews changes
• Plan - gathers context and returns a plan without file writes

• File Picker — finds relevant files
• Code Searcher — grep-style pattern matching
• Researcher — web and docs lookup
• Thinker — works through hard problems
• Editor — writes and modifies code
• Reviewer — catches bugs and style issues
• Commander — runs terminal commands

1. Tree-sitter scans your repo and builds a code map
2. File pickers and searchers find relevant code
3. Thinkers analyze the problem if needed
4. Editors generate changes
5. Reviewers check for issues; commanders run tests

The server is stateless. It streams requests to the active model provider over encrypted connections. Your code stays local; only relevant context is sent.