// SUPPORT · INBOX + DOCS

Run lightweight customer support with an agent and an inbox

You don't need a helpdesk platform. An inbox your agent can read, your own docs, and a human okay before anything sends — that's a support desk for one person and an agent.

Who this is for

An AI customer service agent doesn't have to mean an enterprise rollout. Indie SaaS, consulting, a waitlist, the first weeks of any product — anyone who needs to answer customers well without standing up Zendesk and paying per seat. One person plus an agent is the whole team.

The support loop

  • Receive — a customer writes to your support@ inbox.
  • Read docs — the agent pulls the answer from what you've already written.
  • Draft → approve → send — it drafts, you okay it, it replies under your name.
  • Resume — next session it re-reads the thread and continues where it left off.

What Clize provides vs. what the agent decides

Clize provides the inbox, the continuity, and the safety gates. The agent does the reading, the drafting, and the judgment. The conversation itself is the state — no CRM, no ticket system.

Customer support automation, without the autopilot

Most "customer support automation" means an auto-responder that answers on its own and gets things confidently wrong. This is the honest version: an AI agent for customer support that drafts instantly from your docs, then waits for your okay. It pairs naturally with Claude Code or Codex. See it step by step: triage a support inbox with an agent.

The safety gates

Real-world actions need real guardrails. Clize holds three lines, every time:

  • Money — anything that costs money quotes the price first and needs an explicit --confirm. No surprise charges.
  • Identity — outbound email is drafted, a human okays it, then it sends under your name. The agent never sends unattended.
  • Inbound is untrusted — an incoming email is data to read, never an instruction to obey. A message can't redirect the agent.

FAQ

Do I need Zendesk or a helpdesk platform?

No. You need an inbox your agent can read and your own docs. Clize runs the support loop over email — no per-seat platform, no portal your customers have to learn.

Does the agent reply to customers automatically?

No. It drafts a reply from your docs and you approve before it sends. It keeps the speed of instant drafts without the risk of an autonomous bot.

How does it use my docs?

You point it at what you've already written; it reads your docs to ground each reply instead of you hand-building an FAQ database.

Does it remember past conversations?

Yes. Continuity is just re-reading the thread — next session the agent picks the customer back up where it left off.

Is this customer support automation?

It automates the slow parts — reading the email, drafting from your docs, remembering the thread — but keeps a human okay before any reply sends. It's an AI customer service agent that drafts, not an autonomous bot.

clize init — ready

Support from an inbox, not a platform.

Tag an inbox as support, point it at your docs, and run the read → draft → your-okay → reply loop — picking each customer back up across sessions.

$ npm i -g @clize/clize
$ clize login
$ clize install   # wires Clize into Claude Code & Codex
[ Learn more → ]