Cerebretron Code

Cerebretron's AI coding agent, inside VS Code. It reads your workspace, edits files, runs commands, searches your codebase, and taps Cerebretron's cloud tools (engines, FEA, math, web) — all from a streaming chat in the sidebar or a full editor tab.

Highlights

• Terminal and remote work — open a native workspace terminal, or select terminal output and use Cerebretron: Explain Selected Terminal Output to prepare an editable chat draft. The Remote hub opens existing cloud IDEs, copies account-protected phone links, and guides VS Code Remote Tunnels setup. Cloud workspaces open in the browser. Computer access requires the same GitHub or Microsoft account on both devices and a running tunnel host; save files before switching devices. Remote access shares files and terminals, not the desktop screen or unsaved editor buffers.
• Workspace interface — the website's warm charcoal, cream, and orange palette, crisp SVG controls, saved drafts, keyboard-accessible menus, and a focused layout for the sidebar or editor tab. Light and high-contrast editor themes are supported.
• Review and test — use Cerebretron: Review Current File or Cerebretron: Generate Tests from the Command Palette or editor context menu. Chat starters also help review changes and investigate diagnostics.
• Precise edits — unique, exact search/replace targets preserve line endings. Native reviews recheck the original file and unsaved editor changes before applying; failed saves remain visible with recovery guidance. Ask first mode presents an approval diff; Auto-edit and YOLO use the same file checks.
• Permission modes — Ask first (approve each step), Auto-edit (apply straight to files, review the diff), Plan (investigate → propose a plan before touching code), YOLO (run unattended).
• Effort — Standard (fast), Think (deeper reasoning, streamed live), Max (Pro at maximum reasoning effort), and Ultra (Pro with write → review → audit).
• Live feedback — a running activity indicator with a live token counter, a Writing → Reviewing → Auditing progress stepper for Ultra, and a "Thought for Ns" panel when the model reasons.
• Deep context — CEREBRETRON.md project rules, @file / @git / @problems / @folder / @symbol mentions, language-server navigation, and multi-turn memory. Filenames with spaces use @"src/my component.tsx"; the picker inserts the correct form. Workspace diagnostics include file paths, and omitted or truncated context is called out.
• Checkpoints + Rewind — recorded file edits can be restored while keeping the conversation. Files larger than 1 MB and ambiguous older checkpoints cannot be restored; rewind refuses these rather than guessing. Shell and network side effects are not reversible.
• Sessions — multiple conversations per workspace.
• MCP client — connect your own Model Context Protocol servers; their tools become agent tools.
• Editor-power tools — inline edit (Cmd/Ctrl+K Cmd/Ctrl+I), "Fix with Cerebretron" quick-fixes, ghost-text completions, slash commands.
• Rich chat — syntax-highlighted code, isolated diagram images, and a comfortable reading width in editor tabs. Copy complete replies or tool output, expand results with the keyboard, and edit an earlier request as a draft. Plans are approved for the current conversation only.

Install

From the VSIX: Extensions panel → ⋯ → Install from VSIX… → pick cerebretron-code-<version>.vsix. Or:

code --install-extension cerebretron-code-<version>.vsix

Then reload VS Code and open Cerebretron Code from the Activity Bar.

Sign in

Click Sign In (or run Cerebretron: Sign In) → Sign in with browser — authorize in your browser, no key to copy. (Pasting a gns_ API key from cerebretron.ai still works; it's stored in VS Code SecretStorage, encrypted, never in settings.)

Using it

• Chat: Enter sends · Shift+Enter for a newline · Send → Stop cancels.
• Open in an editor tab: the ⤢ header button, the Cerebretron status-bar item, or Cerebretron: Open Chat in Editor.
• Pick Mode & Effort: the Mode and Effort buttons in the chat bar.
• Reference context: type @ for a file/symbol/git/problems picker.
• Slash commands: type / — /explain /fix /tests /doc /review, /plan, /init (generate CEREBRETRON.md), /model, /usage, /context, /compact, /mcp, /clear, /help, plus your own from .cerebretron/commands/*.md.
• Inline edit: select code → Cmd/Ctrl+K Cmd/Ctrl+I → describe the change → review the diff.
• Attach an image: paste a screenshot or click ＋, then ask about it.

Effort levels

Standard and Think retain your model picker selection. To send an image, use an image-capable model with Standard effort.

Commands & keybindings

• Cerebretron: Sign In / Sign Out
• Cerebretron: New Chat
• Cerebretron: Open Chat in Editor
• Cerebretron: Select / Cycle Permission Mode (Cmd/Ctrl+K Cmd/Ctrl+M)
• Cerebretron: Rewind Checkpoint…
• Cerebretron: Sessions / History…
• Cerebretron: Manage MCP Servers…
• Cerebretron: Edit Selection (inline) (Cmd/Ctrl+K Cmd/Ctrl+I)
• Cerebretron: Explain Selection, Toggle Inline Completions

Settings

How the agent works

Each turn runs a tool loop in the extension host: the model emits tool calls, the extension executes them, feeds the results back, and loops until it answers.

• Read-only tools (read_file, list_dir, codebase_search, LSP nav) run automatically and in parallel.
• Mutating tools (edit_file, write_file, run_shell, run_in_terminal) are gated — you review a diff or the command first (subject to the permission mode).
• Cloud tools (engines, CAD→FEA, math, web search, diagrams, images, knowledge) run server-side via the Cerebretron API.

Privacy & security

• API keys live in VS Code SecretStorage (encrypted), never in settings.
• File access is confined to your workspace (symlink-escape-proof).
• stdio MCP servers run only in a trusted workspace; http MCP is SSRF/DNS-rebind hardened.
• No telemetry is sent unless you explicitly opt in.

See CHANGELOG.md for version history.