Headless Mode & SDK

Working Code

Use the claude -p flag for non-interactive (headless) mode:

claude -p "Analyze the architecture of this project"

The result is printed and the process exits. No interactive session is opened, making it ideal for scripts.

You can also pipe input:

cat error.log | claude -p "Analyze this error log"

For JSON output:

claude -p --output-format json "Analyze the dependencies in package.json"

You get structured JSON that other tools can parse.

Try It Yourself

Build a daily automation script:

#!/bin/bash
# daily-check.sh — Morning project health check script

echo "=== Code Quality Check ==="
claude -p "Find all TODO, FIXME, and HACK comments in src/ and prioritize them"

echo ""
echo "=== Security Check ==="
claude -p "Check if any dependencies in package.json have security vulnerabilities"

echo ""
echo "=== Test Status ==="
claude -p "Run npm test and analyze the cause if anything fails"

Register this script in cron or CI to run automated daily checks.

"Why?" — Beyond Conversations, Into Systems

So far you've used Claude Code interactively in the terminal. Headless mode turns Claude Code into a CLI tool. Use it as one step in a pipeline, just like grep, jq, or curl.

Key Flags

| Flag | Function | | ----------------------------- | --------------------------------------------------- | | -p "prompt" | Non-interactive execution (prints result and exits) | | --output-format json | JSON format output | | --output-format stream-json | Streaming JSON output | | --max-turns N | Limit maximum number of turns |

Pipeline Usage Patterns

# Analyze error logs
cat error.log | claude -p "Find the root cause of this error"

# Review a PR diff
gh pr diff 42 | claude -p "Review these changes as a senior developer"

# Code conversion
cat legacy.js | claude -p "Convert this code to TypeScript" > modern.ts

# Generate release notes
git log --oneline v1.0..HEAD | claude -p "Write release notes from these commits"

Claude Code SDK

Call it programmatically from TypeScript/JavaScript:

import { claude } from "@anthropic-ai/claude-code";

const result = await claude({
  prompt: "Analyze security vulnerabilities in this file",
  workingDirectory: "./my-project",
});

console.log(result.stdout);

With the SDK you can:

• Build custom automation tools.
• Integrate Claude Code into CI/CD pipelines.
• Automate code analysis in monitoring systems.

CI/CD Integration Example

# .github/workflows/ai-review.yml
name: AI Code Review
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Install Claude Code
        run: npm install -g @anthropic-ai/claude-code

      - name: Review PR
        run: |
          gh pr diff ${{ github.event.number }} | \
            claude -p "Review this PR for bugs, security, and performance." \
            > review.md
        env:
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

      - name: Post Review Comment
        run: gh pr comment ${{ github.event.number }} --body-file review.md
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

Deep Dive

result=$(claude -p "Analyze this" 2>/dev/null)