Blog

Guides on cutting AI token usage and shipping frontend fixes faster with AI coding tools.

Codex CLI

How to Reduce Token Usage in Codex CLI

OpenAI's Codex CLI bills reasoning tokens, output verbosity, and prompt caching as three separate, independently-controllable levers. Here's what each config.toml setting actually changes about your bill.

Gemini CLI

How to Reduce Token Usage in Gemini CLI

Gemini CLI's Auto routing and automatic caching quietly move your token spend in opposite directions. Here's how to check which model actually ran, and why chasing manual caching can cost more than it saves.

Claude Code

Fable 5 + Sonnet 5: What Anthropic's Own Benchmark Data Says About Claude Code Costs

Anthropic published the numbers itself: pairing Fable 5 with Sonnet 5 keeps 96% of solo-Fable performance at under half the price. Here's the data and how to set it up.

Claude Code

10 Tips to Stop Burning Your Tokens in Claude Code

Token waste in Claude Code isn't an abstract number — it's a dollar figure. Ten ways to see exactly where that money goes and stop paying for it.

Claude Skills

Using Claude Skills to Reduce Token Usage

Skills only load their full instructions when they're actually relevant. Here's how to structure your own so they save context instead of costing it.

OpenClaw

How to Reduce AI Token Usage in OpenClaw

The same context discipline that works across every AI agent applies to OpenClaw: scope what loads, batch what you send, and skip the screenshots.

OpenCode

How to Reduce Token Usage in OpenCode

OpenCode's provider flexibility is also its biggest token trap. Model choice, session hygiene, and scoped instructions actually move the number.

OpenAI API & MCP

Prompt Engineering to Reduce Token Usage (OpenAI API, MCP, and Beyond)

Prompt caching, tighter system prompts, leaner MCP tool schemas, and fewer few-shot examples — the levers that actually cut token spend at the API level.

Claude Code

9 Ways to Reduce Token Usage in Claude Code

I burn less context per session since fixing these nine habits in Claude Code — from /clear and /compact to swapping screenshots for structured UI context.

Cursor

How to Reduce Token Usage in Cursor

Cursor's context settings, @file targeting, and mode choice all move your token spend. Here's what actually works, plus Cursor's own Dynamic Context Discovery.

GitHub Copilot

How to Reduce Token Usage in GitHub Copilot

Copilot's seat-based pricing hides the token cost, but it still shows up as slower, shallower answers. Here's what actually fixes it.

Windsurf

How to Reduce Token Usage in Windsurf

Windsurf's Memories, Rules, and open-tab context all affect your credit spend. Here's how to use each one deliberately instead of by accident.