Tools Library
Safe-by-default API tools for AI agents: dry-run plans, explicit apply, verification, receipts.
Qwayk tools are small command-line tools (CLIs) for one API each. They’re built for AI-assisted work without “guessing edits”.
I built them because I don’t want an agent to “kind of” do the right thing. I want it to either do the exact thing we planned — or refuse.
If you’re new, don’t overthink it. Pick one tool, run one dry-run, and read the plan like a checklist. If you get stuck, send yourself (or me) the redacted plan/receipt. That’s usually enough to debug fast. If you want a safe first win: run a read-only command first, then stop and review what you got.
The short version: - Dry-run first (plan). - Apply on purpose (writes are gated). - Verify after changes (read-back + receipts).
If you’re new, start with: - Docs - Safety model
Flagship tools
- Ghost API tool — Stable — v0.1.0
- WordPress API tool — Stable — v0.1.0
- Freepik API tool — Stable — v0.1.0
- Plausible API tool — Stable — v0.1.0
Also available
- Mercury API tool — Stable — v0.1.0 (GET-only API layer)
- Statuspage API tool — Stable — v0.1.0
- Pinterest API tool — Stable — v0.1.0
- Unsplash API tool — Stable — v0.1.0
- Amazon PA‑API tool — Stable — v0.1.0
- Google Ads API tool — Beta — v0.1.0 (read-only)
- Meta Ads API tool — Beta — v0.1.0 (read-only)
- Cloudflare API tool — Experimental — v0.7.5
- Dynadot API tool — Experimental — v0.3.0
Status labels (what I mean)
- Stable: safety contract is locked; docs/tests are solid.
- Beta: safe, but still changing quickly.
- Experimental: available, but not guaranteed (expect churn).
Trust & policies
Start here
If you’re new, start with the docs first, then pick a tool.
