98 guides
Run Qwen3-Coder locally with LM Studio on Windows and connect it to OpenCode. Model download, the local server, config, hardware needs, and the common fixes.
Run a DeepSeek model locally with Ollama on Windows and connect it to Claude Code via Claude Code Router. Install, model pull, config, hardware needs, and fixes.
WSL vs native Windows for running AI coding agents like Claude Code, Codex, OpenCode, and Aider. Performance, env vars, git, proxies, and which to choose.
Install and run Aider on Windows using WSL. Python and pipx setup, connecting a cheap model, git workflow, config file, and the WSL-specific gotchas to avoid.
Install and run OpenCode on Windows using WSL. Node setup, install, adding a cheap provider, config location, and the WSL-specific gotchas to avoid.
Install and run OpenAI Codex CLI on Windows with WSL. Setup, config.toml location, connecting a cheap model, env vars, and the WSL-specific gotchas to avoid.
Install and run Claude Code on Windows using WSL. Node setup, install, connecting a cheap model, persisting env vars, and the WSL-specific gotchas to avoid.
Use cc-desktop-switch on Windows to configure DeepSeek, Kimi, GLM, and Bailian providers for Claude Desktop. What it does, setup, the trade-offs, and a simpler path.
Run a LiteLLM gateway on Windows to serve Claude Code, Codex CLI, and OpenCode from one place. Unified keys, spend tracking, model routing, and the common fixes.
Set up claude-code-mux on Windows for fast, fail-over multi-provider routing of Claude Code across DeepSeek, MiniMax, Kimi, and 15+ backends. Install and config.
Set up claude-code-proxy on WSL to run Claude Code against your ChatGPT or Kimi subscription via a local Anthropic-compatible proxy. Install, configure, and fixes.
Use deepclaude on Windows to run Claude Code's agent loop on DeepSeek V4 or any Anthropic-compatible backend. Install, configure, verify, and the trade-offs.
Build a Cursor alternative on Windows with open-source tools and DeepSeek. Cline or OpenCode plus a cheap model give you AI coding without a subscription.
OpenRouter vs direct provider APIs for Claude Code: cost, failover, convenience, and privacy. When the aggregator is worth a small premium and when to go direct.
What Claude Code really costs and practical ways to cut it: cheaper model backends, routing by task, cache and off-peak discounts, and tighter prompting habits.
Free tier AI coding in 2026, honestly assessed: provider free credits, OpenRouter free models, NVIDIA NIM, and local models. What's genuinely useful and what isn't.
Coding plans vs pay-per-token in 2026: how each bills, the break-even point, rolling usage windows, and a simple method to pick the cheaper option for your usage.
The best cheap AI coding plans compared in 2026: Z.AI GLM from $18/month, Alibaba's $50 plan, MiniMax token plans, and MiMo. Limits, links, and which to pick.
The cheapest AI coding APIs in 2026 compared: DeepSeek, GLM, Kimi, MiniMax, MiMo, and OpenRouter routes. Pricing links, pay-per-token vs plans, and how to pick.
Set up OpenCode with StepFun Step 3.7 Flash for a very cheap, multimodal reasoning coding agent. Provider config, model IDs, low pricing, and the common fixes.
Doubao Seed 2.0 Pro vs DeepSeek V4 compared for coding: cost, setup ease, ecosystem, and which cheap model to pick for Claude Code, Codex, or OpenCode.
Run ByteDance Doubao Seed Code with Claude Code via Volcengine Ark and Claude Code Router. API key, endpoint, model choice, pricing, and the common fixes.
Xiaomi MiMo pricing explained: current MiMo V2.5 and V2.5 Pro API rates, cache-hit discounts, no long-context surcharge, token plans, free trial access, and cost estimates.
Configure Codex CLI to use Xiaomi MiMo V2.5 on Windows. config.toml provider, wire_api, big-context coding, current V2.5 API pricing, and the common fixes.
Run Xiaomi MiMo V2.5 Pro with Claude Code for cheap, big-context coding. API key, endpoint setup, model choice, current V2.5 API pricing, free trials, and the fixes.
MiniMax M2.5 vs DeepSeek V4 vs GLM-5 compared for agentic coding in 2026: cost, speed, reasoning, plans vs pay-per-token, and which to pick for your workflow.
Set up OpenCode with MiniMax M2 for cheap agentic coding. Provider config, model IDs, lightning mode, pricing and caching, and the common fixes — Windows or WSL.
Run MiniMax M2.5 with Claude Code for cheap agentic coding. API key, endpoint setup, model choice, pricing and lightning mode, and the common fixes — Windows or WSL.
Kimi K2 pricing explained: Thinking vs Turbo vs K2.5 vs K2.6, cache-hit discounts, provider price differences, and how to estimate your coding cost on Moonshot.
Set up OpenCode with Kimi K2.5 Turbo for fast, cheap agentic coding. Moonshot provider config, model IDs, pricing and cache discounts, and the common fixes.
Configure Codex CLI to use Kimi K2 through Moonshot's OpenAI-compatible endpoint. config.toml provider, wire_api, model choice, pricing, and the common fixes.
Run Kimi K2 Thinking with Claude Code using Moonshot's Anthropic-compatible endpoint — no proxy. API key, environment variables, model choice, pricing, and fixes.
Use GLM models with the Cline extension in VS Code on Windows. OpenAI-compatible provider setup, model choice, plan vs pay-per-token, plan mode, and fixes.
Set up OpenCode with GLM models from Z.AI for a cheap, open-source coding agent. Provider config, model IDs, coding plan vs pay-per-token, and the common fixes.
Configure Codex CLI to use GLM-5 on Windows through Z.AI's OpenAI-compatible endpoint. config.toml provider, wire_api, model choice, coding plan, and fixes.
GLM-4.6 vs GLM-5 vs GLM-5-Turbo compared for coding in 2026: capability, speed, plan availability, and which GLM model to pick for Claude Code and other agents.
Run GLM-5 with Claude Code using Z.AI's Anthropic-compatible endpoint — no proxy. API key, environment variables, model choice, the coding plan, and the fixes.
Use Qwen3-Coder with Aider for cheap, git-native pair programming. DashScope config, model flags, big-context editing, pricing and coding plan, and the fixes.
Run Qwen3-Coder-480B on Cerebras through OpenCode for extremely fast agentic coding. Provider config, API key, model ID, why Cerebras is fast, pricing, and fixes.
Qwen3 Max API pricing explained, plus how to use it cheaper: cache discounts, the Alibaba coding plan, and when qwen3-coder-plus or qwen3.5-plus is the better buy.
Install and run Qwen Code CLI on Windows. Node setup, API key from DashScope, model config, first task, pricing and coding plan, and the common fixes.
Run Qwen3-Coder with Claude Code on Windows via Claude Code Router and Alibaba's DashScope. API key, config, big-context coding, pricing and coding plan, and fixes.
Use DeepSeek with Claude Desktop on Windows with a provider-switch tool. How the desktop app differs from the CLI, setup steps, model choice, and the trade-offs.
DeepSeek V4 pricing explained with current V4 Flash and V4 Pro API rates, cache-hit discounts, Anthropic/OpenAI base URLs, and how to estimate coding cost.
Use DeepSeek with the Cline extension in VS Code for cheap, in-editor AI coding. OpenAI-compatible provider setup, model choice, plan mode, pricing, and fixes.
Set up Aider with DeepSeek on Windows for git-native AI pair programming at pennies per session. Install, API key, model flags, config file, and fixes.
Set up OpenCode with DeepSeek for a cheap, open-source coding agent. The opencode.json provider block, model IDs, API key, pricing, and fixes — on Windows or WSL.
Configure OpenAI Codex CLI to use DeepSeek through config.toml. The custom provider block, wire_api gotcha, model choice, pricing, and fixes — on Windows or WSL.
Run DeepSeek V4 with Claude Code inside WSL on Windows. Node setup, the Anthropic-compatible endpoint, persistent env vars, model choice, and pricing notes.
Run DeepSeek V4 with Claude Code on Windows using its Anthropic-compatible endpoint. API key, environment variables, model choice, pricing, and the common fixes.
Run Claude Code on any backend using Anthropic-compatible endpoints. How ANTHROPIC_BASE_URL works, which providers support it, and setup on Windows — no proxy.
Claude Code vs Codex CLI vs OpenCode vs Aider compared for Windows in 2026: workflow, custom model support, cost, and which terminal coding agent to pick.
The real security risks of MCP servers in Claude Code — over-broad access, secrets in config, prompt injection, untrusted servers — and how to lock each one down.
MCP servers cost context tokens before you type. Why each connected server's tools eat your context window, how to measure it, and how to fix it in Claude Code.
The best free and cheap Claude Code alternatives for Windows in 2026: Codex CLI, OpenCode, Aider, Cline, Kilo Code, and more, with download links and honest trade-offs.
Fix Claude Code MCP servers that won't connect — the Windows cmd /c npx wrapper, PATH issues, bad commands, auth failures, and stdout corruption. Step-by-step.
Connect the GitHub MCP server to Claude Code on Windows and WSL. Token auth, what it unlocks — issues, PRs, code search — example workflows, and scope cautions.
Add the Playwright MCP server so Claude Code can drive a real browser — navigate, click, screenshot, assert. The verify-my-change loop, headless vs headed, WSL notes.
Connect Postgres to Claude Code with an MCP server the safe way — read-only roles, non-prod targets, schema inspection, and query workflows on Windows and WSL.
Configure OpenCode to use DeepSeek, GLM, Kimi, or Qwen through custom providers. opencode.json with the OpenAI-compatible SDK, model IDs, API keys, and fixes.
Use Claude Code with an MCP server to inspect containers, read logs, and debug Docker on Windows and WSL — what to connect, safe scope, and example workflows.
Set up Claude Code for IT and infrastructure work — install on Windows and WSL, user vs project config, safe permissions, the right MCP servers, and guardrails.
Compare in-session Claude Code subagents with running parallel agents across git worktrees. Trade-offs, when each fits, and a worktree setup for Windows and WSL.
MCP servers and plugins solve overlapping problems in Claude Code. When to add a raw MCP server, when a plugin makes sense, and how to keep your context lean.
Configure OpenAI Codex CLI to use DeepSeek, Qwen, GLM, or Kimi through config.toml. Custom model_providers, base URLs, env keys, and the wire-protocol gotcha explained.
Claude Skills, Claude Code, and Projects get confused constantly. What each one is, where they overlap, and which to use for repeatable work versus coding.
Step-by-step guide to building a custom Claude Code subagent: write the frontmatter, a focused system prompt, scope tools tightly, test delegation, and iterate.
Build your own Claude Code Skill from scratch — create a SKILL.md, write a description that triggers it, structure the instructions, add scripts, test it, and share it.
Plugins, skills, and MCP servers sit at different layers in Claude Code. What each is, how they nest, and a clear guide to which one you actually need.
Install Claude Code Router on WSL and route Claude Code to DeepSeek, GLM, Qwen, or Kimi. Node setup, config.json, env vars, and the WSL-specific gotchas to avoid.
CLAUDE.md gives Claude Code persistent project context. Where it lives, what to put in it, the user vs project files, imports, and how to keep it lean and useful.
Set up a custom Claude Code status line on Windows and WSL — show model, context, cost, and git branch. The settings config, a working script, and the gotchas.
Build your own Claude Code slash commands — markdown prompt files in .claude/commands, arguments, namespacing, and when a command beats a skill. With examples.
Install Claude Code plugins step by step — add a marketplace, browse with /plugin, install and manage plugins, and build your own marketplace to share a setup.
Install and configure Claude Code Router on Windows to route Claude Code to DeepSeek, GLM, Qwen, Kimi, and more. Full config.json, per-task routing, and fixes.
The Claude Code status line setups worth copying in 2026 — minimal, cost-aware, git-focused, and powerline styles — plus what each field is good for.
The MCP servers worth adding to Claude Code for ops work: filesystem, GitHub, Postgres, SQLite, Docker, fetch, Sentry, and Playwright — with honest caveats.
The built-in and custom Claude Code slash commands worth knowing in 2026 — context, cost, model switching, clear, plus custom commands you should build.
Run Claude Code on cheaper models like DeepSeek, GLM, Kimi, and Qwen. The three connection methods, pricing and coding plans compared, and how to set it up on Windows.
Practical Claude Code hook examples worth copying — auto-format, block destructive commands, notify on stop, log activity, and load context on session start.
The Claude Code plugins actually worth checking in 2026, with install commands, source links, and why GitHub, Playwright, LSP, security, review, docs, and workflow plugins make the cut.
The most useful Claude Code subagent patterns for 2026: code reviewer, test writer, debugger, security auditor, docs writer, refactorer, and explorer — with tools and models.
OpenClaw and Claude Code both write code with AI, but they're built for different jobs. What each is best at, where they overlap, and how to choose — or use both.
Use OpenClaw's coding-agent skill to read issues, write fixes, run tests, and open pull requests on its own. How it works, how to set it up, and how to keep it safe.
The Claude Code Skills worth using in 2026: bundled /code-review, /debug, /run, /verify, plus installable skill plugins with links and install commands.
Skills, MCP servers, and subagents solve different Claude Code problems. What each does, what it costs, and a clear decision guide for picking the right one.
Set up OpenClaw with GLM from Z.AI for a flat-rate self-hosted assistant. The config, the coding plan vs pay-per-token choice, key handling, and the gotchas.
Use Claude Code hooks to run shell commands on agent events — auto-format after edits, block risky commands, log activity. The events, config, and working examples.
Point OpenClaw at DeepSeek for a self-hosted AI assistant that costs cents to run. The config, the OpenAI-compatible endpoint, key handling, and the gotchas.
OpenClaw is the viral open-source personal AI assistant that runs on your own machine and your own API keys. What it does, how it works, and why it took off.
The MCP servers actually worth installing in Claude Code in 2026 — GitHub, filesystem, Postgres, Playwright, fetch, Sentry, Slack, and memory. With add commands.
Learn how Claude Code subagents work: separate context windows, custom system prompts, scoped tools, and per-agent model routing. Full example agent file included.
Pillar Guide
A map of OpenClaw — how to install it, point it at cheap models like DeepSeek and GLM, use skills and the coding agent, and pick between the popular forks.
Add MCP servers to Claude Code on Windows and WSL: claude mcp add, scopes, .mcp.json, verifying with /mcp, and the Windows cmd /c npx wrapper gotcha explained.
Build a minimal MCP server from scratch with the official TypeScript SDK, expose one tool over stdio, connect it to Claude Code, and test it end to end.
What Claude Code Skills are, the SKILL.md format, where they live, how Claude loads them on demand, and a worked example you can drop in and use today.
A plain-English guide to the Model Context Protocol (MCP): the client/server model, tools vs resources vs prompts, transports, and why it matters for Claude Code.
Pillar Guide
A map of Claude Code's advanced features — MCP servers, Skills, subagents, hooks, plugins, slash commands, status line, and CLAUDE.md — and when to use each.