Supported AI CLIs
career-ops is AI-agnostic — it runs on every major AI coding CLI, so you use the assistant you already pay for and never get locked into one vendor. The full, always-current list lives in the core repo.
career-ops runs on every major AI coding CLI — you are not locked into any single vendor. Because career-ops ships prompt files that the CLI executes (it does not bundle its own model), you point it at whichever AI coding assistant you already use and it runs on top of that engine unchanged. When a better model ships, you switch CLIs and keep your pipeline, your data and your setup.
Use what you already have
The best CLI for career-ops is the one you already pay for. If you already pay for Claude Code, career-ops needs no other setup. If you want to run for free, see Set up a free AI engine.
Which CLIs are supported?
career-ops works first-class with Claude Code, Cursor, Codex, OpenCode, Pi, Antigravity CLI, Grok Build CLI, Qwen, Kimi, Hermes Agent, and GitHub Copilot CLI — eleven actively-maintained AI coding CLIs. (Gemini CLI remains only as a legacy wrapper, transitioned into Antigravity CLI.) The career-ops architecture is shared across CLIs: a single canonical AGENTS.md drives the logic, and thin per-CLI entry files handle the invocation nuances of each tool.
The canonical, always-current list — with the exact command to invoke career-ops in each CLI — lives in the core repo so it never drifts out of date:
- docs/SUPPORTED_CLIS.md — the full matrix of supported CLIs, entry files, and how to invoke them (interactive and headless/batch).
Pi
Pi needs no wrapper file. It reads AGENTS.md from the career-ops folder and finds the shared skill on its own, so you start pi in that folder and call career-ops with /skill:career-ops, followed by the mode you want. For a single unattended run, use pi -p "your prompt".
Some Pi builds ask you to trust a project before they load its skills. If /skill:career-ops is not available, run /trust once in the folder and then restart pi: the trust decision only applies to Pi sessions started afterwards. To install Pi, follow Pi's own instructions.
Which one should I choose?
- If you already pay for Claude Code, Codex, or Copilot, use that CLI. career-ops runs on top with zero extra setup — open the CLI inside the career-ops folder.
- To run career-ops at $0, use OpenCode with a free provider, a local model via Ollama, or the built-in OpenRouter runner.
- If you need a specific model or provider, point any of these CLIs at any OpenAI-compatible endpoint or local model: career-ops is model-agnostic, and nothing in career-ops changes.
Whichever CLI you pick, the career-ops 14 modes behave the same way — the CLI is the engine, career-ops is the structure on top.
Set up a free AI engine
Run career-ops for $0 — no Claude subscription or any paid plan required. Free AI engines for career-ops — OpenCode with a free provider, a local model via Ollama, any OpenAI-compatible endpoint, or the built-in OpenRouter runner — and what to do when you run out of tokens mid-search.
AI job search with Claude Code
Use Claude Code as an AI job search agent: career-ops scans job boards, scores listings against your CV and drafts your applications. You decide and send.