Claude's Replies Longer Than the Fix? Switch Claude Code to the Concise Output Style.

At a glance

• One command, /output-style concise, makes Claude lead with the result and drop the lead-in, narration and closing recap.
• Switch to Explanatory when you are new to a codebase and want short Insight notes on why Claude made each choice.
• Save a default per project or for every project, and spell the setting value exactly right.

Here is an example of the problem. Say you ask Claude to rename a config key. It does, in one edit, and then writes four paragraphs: what it was about to do, what it did, which files it checked and a recap of all three. You scroll past them to find the one line you needed. Do that thirty times a day and you start skimming, which is how you miss the line that actually matters. Claude Code ships a built-in style for exactly this.

Why it matters

An output style is a set of instructions Claude Code sends with every request, so you set it once instead of typing "keep it short" in each prompt. The Concise style puts the answer in the first sentence and answers simple questions in one to three sentences. The docs say it does the engineering work as thoroughly as Default, and it still writes in full when you ask for detail. Error reports, failing test output, security warnings and confirmations for destructive actions also keep their full content. Shorter replies also mean fewer output tokens. The move: turn on Concise for routine work and keep Explanatory for the days you are learning a repo.

How to switch your output style

You need Claude Code v2.1.269 or later for the /output-style command. We built these steps from Anthropic's docs and have not run them in a live session.

1. Check where you are. Run /output-style with no argument. It lists the styles you can pick and marks the current one.
2. Switch to Concise. Run /output-style concise. The command ignores case, and Claude uses the new style starting with your next message, with no need to /clear.
3. Prefer a menu? In the terminal, run /config and select Output style. In the VS Code extension, type / in the prompt box and pick Output styles.
4. Know where the choice lands. The command and both menus save it to .claude/settings.local.json, so it applies to this project on your machine only.
5. Make it your default everywhere. Add "outputStyle": "Concise" to ~/.claude/settings.json. A project's own settings still win over that value.
6. Use Explanatory or Learning when you want the reasoning. Explanatory adds two or three point Insight blocks in the conversation, not in your files. Learning adds those and leaves small TODO(human) pieces for you to write.
7. On the Desktop app, set the field in a settings file such as .claude/settings.local.json, because /config there opens Settings instead of a menu.

/output-style                      # list styles, current one marked
/output-style concise              # switch, applies from your next message
/output-style explanatory          # add Insight notes on each choice
/output-style default              # back to standard behavior
/config                            # menu: Output style

# ~/.claude/settings.json (all projects) or .claude/settings.local.json
{ "outputStyle": "Concise" }

Gotchas

• The settings value is case-sensitive. In a settings file, write Proactive, Concise, Explanatory or Learning exactly. A value like concise there quietly gives you Default. Only the slash command ignores case.
• A style is an instruction, not a guarantee. If something must happen every time, such as formatting after each edit, use a hook. For project facts, use CLAUDE.md, which stays loaded whichever style you pick.
• Subagents keep their own voice. Styles apply to the main conversation and to forks, but other subagents run their own system prompt.
• Explanatory and Learning cost more output. They produce longer replies by design. Concise does the opposite.
• Switching mid-session keeps the cache. Claude Code delivers the new style as a message, so the switch still reads the system prompt and earlier conversation from the prompt cache.

Sources:

• Output styles, Claude Code docs
• Prompt caching, Changing output style, Claude Code docs
• Claude Code changelog, v2.1.237 Concise style and v2.1.269 /output-style

Recommended AI prompt

Copy this paragraph into ChatGPT, Claude, Gemini, Grok, or whatever AI you use.