
Claude Code with Ollama: Run Local Models on Your Machine
How to run Claude Code with Ollama for local AI models offline: the ANTHROPIC_BASE_URL setup, proxy and router options, the trade-offs, and when it's worth it.
Category
Hands-on guides to Claude Code — Anthropic's terminal coding agent — from install and first run to MCP, subagents, hooks, and the errors you'll actually hit.
Claude Code is the agent we run all day: this entire site is built and maintained in it, so the guides in this category are written from the driver's seat. That means the workflow advice, the model-routing habits, and the troubleshooting come from real sessions on a real project, not a feature tour. Where a fix only reveals itself after you've hit it a dozen times, that's the kind of detail we try to capture here.
Start with what Claude Code is and how to install it, then move into the pieces that make it powerful — MCP servers, subagents, memory files, permissions, and headless automation — and keep the error guides handy for when a run stalls. Whether you're deciding if it's worth a subscription or wiring it into CI, the goal of this collection is to shorten the gap between installed and genuinely productive.

How to run Claude Code with Ollama for local AI models offline: the ANTHROPIC_BASE_URL setup, proxy and router options, the trade-offs, and when it's worth it.

Learn how to install claude-mem, the community tool that gives Claude Code persistent memory across sessions, plus how it compares to built-in options.

Getting 'readablestream is not defined' from the Claude SDK in Node? Learn the root cause and the ordered fixes: upgrade Node, polyfill, or configure your runtime.

Learn how to stop Claude Code asking for permission the safe way — session approvals, allowlists in settings.json, trusted folders, and the risky skip flag.

Claude skills vs commands explained: skills load automatically when Claude decides they fit, slash commands run when you type them. See when to build each.

Practical Claude Code tips and tricks: manage context with CLAUDE.md, use plan mode and subagents, run headless jobs, and learn the commands that speed up work.

What claude superpowers is: a popular open-source collection of Claude Code skills that adds ready-made workflows. How to install it safely and use it well.

The Claude Code statusline puts model, context and cost in your terminal. Setup in settings.json, custom scripts and recipes worth copying.

Node.js for Claude Code: the native installer needs none, npm wants Node 22+. Check your version, set up nvm per OS, and fix common install errors.

OpenCode is open-source and model-agnostic; Claude Code is Anthropic's polished, Claude-native agent. Feature matrix, cost, and a clear pick for each use case.

Is Claude Code free? The CLI is free to install, but running it needs a paid plan or API credits. What each route costs and the cheapest legit ways to use it.

Run claude update to get the latest Claude Code in seconds. Full steps for the native installer, npm, and package managers — plus how to control auto-updates.

How to change model in Claude Code: type /model to switch between Opus, Sonnet, and Haiku instantly, or set a permanent default with --model.

Codex vs Claude Code: delegation in OpenAI's cloud versus collaboration in your terminal. Benchmarks, costs and which agent fits you.

Everything Claude Code does in one place: install, the core workflow, models, sessions, cost, extensions, remote control, and the errors you'll hit — with deep links.

A 'Claude planner' is Plan Mode in Claude Code: read-only research that ends in a step-by-step Markdown plan you approve before any edits happen.

Cursor is IDE-first — you drive, AI assists. Claude Code is agent-first — you direct, AI drives. Feature table, pricing, and when to use which (or both).

Claude Code hooks are shell commands that fire on lifecycle events — before/after tool use, session start, stop — to format code, run tests, and block risky actions.

Claude Code is Anthropic's agentic coding tool for terminal, IDE, desktop, and web. What it does, what it costs, how to install it, and how to work with it.

Point Claude Code at OpenRouter with two environment variables. Setup steps, honest cost math vs the direct Anthropic API, and the caveats tutorials skip.

Claude Code VS Code extension guide: install from the marketplace, delegate from the sidebar, review inline diffs — and when the CLI wins.

Claude Code subagents are isolated Claude instances that work in parallel with their own context. When to use them, how to set one up, and the cost trade-offs.

Claude Code skills are SKILL.md folders that teach the agent your workflows: how they load, the best ones to install, and how to build your own.

Is the Claude Code source code open source? No — it's proprietary to Anthropic. Here's what the official GitHub repo actually holds, the 2026 sourcemap leak, and how to build on it safely.

Claude Code saves every session automatically. Resume the latest with --continue, pick an older one with --resume, and find them as JSONL transcripts. Full guide.

The rate limit reached Claude Code error hides five limits. Learn which one fired — plan window or API cap — and six fixes that get you working again.

Claude Code Router proxies the agent's requests to DeepSeek, Gemini, OpenRouter or Ollama. Setup, routing rules and the honest trade-offs.

Claude Code Remote Control lets you steer a local session from your phone or browser while it keeps running on your machine. How to enable it, use it, and stay secure.

Claude Code plan mode is a read-only mode where Claude researches and writes a plan before touching your code. Here's how to enter it, approve a plan, and set it as default.

Claude Code is included in Pro and Max and shares one usage pool. Max 5x gives 5x the usage, Max 20x gives 20x. Here's the Pro vs Max math and who should upgrade.

Claude Code MCP connects the agent to your tools, databases, and APIs through the Model Context Protocol. How MCP servers work, how to add one, and security.

Claude Code process exited with code 1? The crash is generic — the fix isn't. Read the real error, fix auth, update Node 18+, reinstall clean.

The frontend design plugin is Anthropic's official Claude Code plugin for distinctive UI. What it changes, how to install it, and how to brief it properly.

Install the Claude Code Get Shit Done (GSD) plugin from the plugin marketplace, then run /gsd:init to start a structured plan-execute-verify workflow. Full setup.

The Claude Code desktop app puts the agent in a visual cockpit: setup on Windows and macOS, the Code tab, and when to pick it over the CLI.

Claude Code API error 500 means Anthropic's side erred, not your key. Status check, retry, update and fresh-session fixes that actually work.

Claude Code is included free in Pro and Max plans, or billed per token on the API. Here's the real cost math, per-day spend data, and how to keep the bill low.

Claude Code CLI documentation in one page: install, core commands, flags, slash commands, settings.json hierarchy and a practical cheatsheet.

Claude Code auto compact summarizes your session when context runs low. Learn what survives, how to read the indicator, and how to control or disable it.

Claude automode lets Claude Code edit and run without per-step approval. How the safety layers work, when to trust it, and how to toggle it.

Connect Blender to Claude Code with one command: claude mcp add blender uvx blender-mcp. Install the addon, wire up the MCP server, and build 3D scenes by prompt.

Authenticate Claude Code in the terminal: the /login browser handshake, the API-key route, headless CI tokens and fixes for auth failures.