Skip to main content
The CLI is designed to be set up by an agent. Paste the block below into Claude Code, Codex, OpenClaw, or any agent that runs shell commands, and it will handle the rest:

Why this works well with agents

  • JSON output for every command (--json). Agents prefer structured JSON over text parsing.
  • Stable error codes in the JSON envelope (error.code). The agent can branch on them without text heuristics.
  • Multi-profile via ~/.zapsterapi/credentials. Tokens never have to live in env vars or flags.
  • Privacy-safe telemetry that helps us improve the CLI without exposing user data.

Copy-paste shortcuts for agents

  • llms.txt (download): full docs index in llmstxt.org format, ready to drop in as initial context for any agent.
  • SKILL.md (copy-paste in the docs): a Claude Code-format skill — paste it at ~/.claude/skills/zapsterapi-cli/SKILL.md and the agent picks up the CLI conventions automatically.