The simple /btw command that cuts Claude Code costs in half

You're deep in a coding session with Claude Code, working through a complex refactoring task. Halfway through, you realize you need to ask about a specific function you spotted earlier. But starting a new conversation would mean losing all that precious context — and potentially doubling your token costs.

That's exactly the problem /btw solves. This new feature lets you ask quick questions about your current work without adding to the conversation history.

What makes /btw different from regular Claude Code chats?

The /btw command lets you ask a quick question about your current work without adding to the conversation history. Think of it as a temporary side channel — you get your answer, but the question and response don't clutter up your main conversation.

Side questions have full visibility into the current conversation, so you can ask about code Claude has already read, decisions it made earlier, or anything else from the session. /btw is the inverse of a subagent: it sees your full conversation but has no tools, while a subagent has full tools but starts with an empty context.

The key difference: Use /btw to ask about what Claude already knows from this session; use a subagent to go find out something new.

How do you use /btw in Claude Code?

Using /btw is straightforward, but there's one crucial detail to get right. The /btw command acts as an isolated sidebar. When you use it, Claude spawns a temporary "ephemeral" agent to answer your query.

Here's how it works:

1. Type /btw followed by your question in the same message
2. Don't just type /btw and press enter — that won't work
3. The question and its answer are not saved to your current session's conversation history
4. This mode runs in read-only mode, doesn't trigger tools, and won't save to the main chat history—once the user closes the window, those related answers will vanish automatically

For example, if you're working on a Python script and need quick clarification about a function:

/btw What does the calculate_metrics function return in this context?

or

/btw Can you explain this regex pattern without changing the main conversation?

Why does /btw matter for token costs?

Token management is one of the biggest pain points in AI coding. Before /btw, interruptions reduce efficiency; /btw mitigates this by allowing seamless side queries.

Here's the token math that matters:

• Before /btw: Every follow-up question adds to your main conversation, accumulating context and driving up costs
• After /btw: The suggestion runs as a background request that reuses the parent conversation's prompt cache, so the additional cost is minimal

This becomes especially valuable during long coding sessions. Claude Code already supports code generation, debugging, and optimization, but /btw enhances multitasking capabilities, potentially reducing development time by up to 20 percent based on similar features in competing tools.

When should you use /btw instead of regular prompts?

This is useful when you want a fast answer but don't want to clutter the main context or derail Claude from a long-running task.

Use /btw for:

• Quick clarifications about existing code
• Asking about decisions Claude made earlier
• Checking syntax or function signatures
• Getting explanations without breaking your workflow

Stick to regular prompts for:

• Requests that require tool usage
• Changes you want to track in conversation history
• Complex questions that might lead to follow-ups
• Anything you might reference later

The rule of thumb: if it's a "by the way" type of question that doesn't need to be part of your permanent record, use /btw.

How do you get started with Claude Code and /btw?

To use the /btw feature, you'll need Claude Code installed and updated to the latest version. In the latest Claude Code 2.1.72 update (released March 10, 2026), Anthropic introduced the /btw (By The Way) command.

What do you need to run Claude Code?

Claude Code requires a Pro, Max, Teams, Enterprise, or Console account. The free Claude.ai plan does not include Claude Code access. You can also use Claude Code with a third-party API provider like Amazon Bedrock, Google Vertex AI, or Microsoft Foundry.

Get started with Claude Code using their official installation guide. The fastest way is the native installer, which works on macOS, Windows, and Linux without requiring Node.js.

How do you install Claude Code?

1. Windows: Open PowerShell and run:

   irm https://claude.ai/install.ps1 | iex
   

2. macOS/Linux: Run in terminal:

   curl -fsSL https://claude.ai/install.sh | bash
   

3. Alternative: Install via package managers like Homebrew (brew install claude-code) or npm (npm install -g @anthropic-ai/claude-code)

After installation, log in by running claude and following the browser prompts. See Authentication for all account types and team setup options.

What about the VS Code extension?

If you prefer working in VS Code, you can install the Claude Code extension from the VS Code Marketplace. The VS Code extension provides a native graphical interface for Claude Code, integrated directly into your IDE. This is the recommended way to use Claude Code in VS Code.

With the extension, you can review and edit Claude's plans before accepting them, auto-accept edits as they're made, @-mention files with specific line ranges from your selection, access conversation history, and open multiple conversations in separate tabs or windows.

What are the limitations of /btw?

While /btw is incredibly useful, it has some constraints by design:

• Read-only mode: This mode runs in read-only mode, doesn't trigger tools
• No persistence: Answers disappear when you close the /btw window
• Limited to context: It can only work with information already in your current session
• No tool access: Unlike regular Claude Code interactions, /btw can't run commands or edit files

These limitations are actually features — they keep /btw lightweight and prevent it from accidentally interfering with your main workflow.

Which other features work well with /btw?

Claude Code has several complementary features that work great alongside /btw:

Subagents handle tasks that require fresh context or external research. On paper, custom subagents are Claude Code's most powerful feature for context management. The pitch is simple: a complex task requires X tokens of input context, accumulates Y tokens of working context, and produces a Z token answer.

Custom slash commands let you create reusable workflows. Custom skills and slash commands let you package reusable workflows in SKILL.md files. You can also build custom agents for specialized tasks. Create a /deploy-staging command, a /review-pr checklist, or a /generate-tests routine that your whole team shares. Claude loads these automatically based on context.

Background tasks handle long-running operations without blocking your main conversation.

The /btw feature represents a shift toward more nuanced interaction patterns with AI coding assistants. Instead of treating every question as equally important, you can now distinguish between workflow-critical queries and casual clarifications.

This seemingly simple addition addresses a real problem that many developers face: the tension between staying in flow and managing AI conversation costs. With /btw, you don't have to choose between context and economy — you can have both.