Claude Code Config Generator
Build a working Claude Code setup in seconds: settings.json, CLAUDE.md, and .mcp.json with the right model, permission mode, and MCP servers — plus an optional private in-browser AI that writes your project context file.
Generate a complete Claude Code setup
Claude Code is configured through a few small files: settings.json (model, permissions, behavior), CLAUDE.md (the project context loaded into every session), and .mcp.json (Model Context Protocol servers). This tool builds all three from a short form, with copy and download buttons for each — no syntax to memorize and no invalid JSON.
Install Claude Code
npm install -g @anthropic-ai/claude-codeOr with the install script:
curl -fsSL https://claude.ai/install.sh | bashWhat each file controls
- settings.json — the default model (Opus, Sonnet, Haiku, or Opus-in-plan-mode), the permission mode, allow/deny rules, transcript retention, and privacy hardening (disable telemetry, error reporting, and auto-update).
- CLAUDE.md — a short, high-signal description of your project: stack, commands, conventions, and guardrails. Claude Code reads it at the start of every session.
- .mcp.json — connects MCP servers such as Filesystem, GitHub, Context7, Playwright, Postgres, and Brave Search so the agent can use external tools.
Permission modes explained
- Default — Claude asks before editing files or running commands.
- Accept edits — file edits apply automatically; commands still prompt.
- Plan — a read-only mode that produces a plan before any changes.
- Bypass permissions — no prompts at all; only use in a sandbox you trust.
Allow and deny rules
Allow rules let trusted actions run without prompting (for example Bash(git:*) or Edit). Deny rules block sensitive paths and dangerous commands no matter what — common picks are Read(./.env), Read(./secrets/**), Bash(rm:*), and Bash(sudo:*).
Private, in-browser AI for CLAUDE.md
The optional AI assist writes a tailored CLAUDE.md from a plain-English description of your project. It runs entirely in your browser using WebGPU — your description is never uploaded, not even to us. A ready-made template is always available without it.
Frequently asked questions
Where do these files go?
Put settings.json in .claude/settings.json inside your project (or ~/.claude/settings.json for global defaults). CLAUDE.md and .mcp.json live in your project root and can be committed so the whole team shares them.
Is this official?
No — this is a free community tool that produces standard Claude Code configuration files. See the official documentation for the full settings reference.
Does it send my data anywhere?
No. Everything is generated locally in your browser, and the optional AI context writer also runs locally via WebGPU.
Generate a complete Claude Code setup
Claude Code is configured through a few small files: settings.json (model, permissions, behavior), CLAUDE.md (the project context loaded into every session), and .mcp.json (Model Context Protocol servers). This tool builds all three from a short form, with copy and download buttons for each — no syntax to memorize and no invalid JSON.
Install Claude Code
npm install -g @anthropic-ai/claude-codeOr with the install script:
curl -fsSL https://claude.ai/install.sh | bashWhat each file controls
- settings.json — the default model (Opus, Sonnet, Haiku, or Opus-in-plan-mode), the permission mode, allow/deny rules, transcript retention, and privacy hardening (disable telemetry, error reporting, and auto-update).
- CLAUDE.md — a short, high-signal description of your project: stack, commands, conventions, and guardrails. Claude Code reads it at the start of every session.
- .mcp.json — connects MCP servers such as Filesystem, GitHub, Context7, Playwright, Postgres, and Brave Search so the agent can use external tools.
Permission modes explained
- Default — Claude asks before editing files or running commands.
- Accept edits — file edits apply automatically; commands still prompt.
- Plan — a read-only mode that produces a plan before any changes.
- Bypass permissions — no prompts at all; only use in a sandbox you trust.
Allow and deny rules
Allow rules let trusted actions run without prompting (for example Bash(git:*) or Edit). Deny rules block sensitive paths and dangerous commands no matter what — common picks are Read(./.env), Read(./secrets/**), Bash(rm:*), and Bash(sudo:*).
Private, in-browser AI for CLAUDE.md
The optional AI assist writes a tailored CLAUDE.md from a plain-English description of your project. It runs entirely in your browser using WebGPU — your description is never uploaded, not even to us. A ready-made template is always available without it.
Frequently asked questions
Where do these files go?
Put settings.json in .claude/settings.json inside your project (or ~/.claude/settings.json for global defaults). CLAUDE.md and .mcp.json live in your project root and can be committed so the whole team shares them.
Is this official?
No — this is a free community tool that produces standard Claude Code configuration files. See the official documentation for the full settings reference.
Does it send my data anywhere?
No. Everything is generated locally in your browser, and the optional AI context writer also runs locally via WebGPU.
You build the idea. I'll ship the product.
Productized MVP development for founders. 9 SaaS apps shipped — yours could be next, in 6 weeks. Secure by default.
ℹ️ Disclaimer
This tool is provided for informational and educational purposes only. All processing happens entirely in your browser - no data is sent to or stored on our servers. While we strive for accuracy, we make no warranties about the completeness or reliability of results. Use at your own discretion.