$ capsul guides
Guides
How token usage actually works in a CLI coding agent, and what brings it down. No sign-up, no gate, no email.
- How to reduce Claude Code token usageSeven concrete ways to cut the tokens Claude Code sends on every turn, from scoping the request to capping the context, with what each one is actually worth.
- Claude Code usage limits, and how to stop hitting themWhat the Claude Pro and Max session and weekly limits actually count, why they arrive mid-task, and the habits that keep a subscription lasting the whole week.
- Codex CLI token usage: reading it, then lowering itWhere the tokens go in an OpenAI Codex CLI session, how to see the real input count before spending a request, and what reduces it without changing the answer.
- Context engineering for coding agentsWhy what you send matters more than what you ask, how context degrades as it grows, and the rules for building context an agent can answer from.
- What an AI coding agent really costs per developerSubscription against API pricing, what separates two developers on the same plan, and how to work out what a team of fifteen actually spends.
- capsul quickstart: install, first run, first numbersInstall capsul with npm, check your environment, open an interactive session, and run a real task with the agent CLI you already use.
- Claude Code /compact vs /clear, and what each one costsWhat /compact and /clear each do to Claude Code's context window, what each costs in tokens, when auto-compaction runs, and a simple rule for choosing.
- Claude Code subagents: what they save, and what they costSubagents keep your main context small, but each one pays its own setup and reads. When they save tokens, when they waste them, and how to see what they used.
- Claude Code vs Codex CLI: where each one spends its tokensHow Claude Code and Codex CLI spend input tokens, what each one reports, how each bills, and how to read a measured benchmark of both agents run on their own.
- What a CLAUDE.md costs in tokens, and how long it should beHow Claude Code loads CLAUDE.md and Codex loads AGENTS.md, what they add to every request even with caching, how to see it, and what to cut or move.
- Claude Opus vs Sonnet vs Haiku: what a coding task really costsClaude list prices per model as of September 2026, the measured cost of the same coding questions on each, and why cheap per token is not always cheap per task.
- What MCP servers cost in tokens, and how to cut itWhat an MCP server adds to every Claude Code and Codex CLI request, from tool definitions to results and cache misses, how to see it, and how to cut it.
- Prompt caching for coding agents: what it saves and what breaks itHow Claude's prompt cache works in a coding agent: what writes and reads cost, 5-minute vs 1-hour lifetimes, what breaks it, and how to read your hit rate.