Proje vitrini hazırlanıyorPreparing project showcaseПодготавливаем витрину проекта

Integrations

API keys & secure access management

How small businesses can manage API keys safely: separate keys, narrow scopes, regular rotation, and what to do in the first five minutes of a leak.

Rocketly · 2026-07-19

In a small business, the trouble usually starts quietly. You hire an outside developer to set up an integration, and they send a perfectly reasonable message: "Can you paste me your API key?" You copy it into the WhatsApp group. The job gets done, the developer moves on — but the key still works, still sits in that chat, and nobody remembers who can see it. API key security is mostly about these quiet risks: not a distant hacker, but forgotten, shared, never-rotated keys.

This article explains how to manage API keys safely even without a technical team: what a key actually is, how to narrow what it can do, when to change it, and what to do in the first five minutes if one leaks.

What an API key actually is

Think of an API key as a password that lets two pieces of software tell each other "I'm allowed to be here." When your website sends a new contact to your CRM, or your accounting tool pulls an invoice, no human is in the loop — the two systems talk through that key.

One distinction sets the boundary of this article. Single sign-on (SSO) and identity management is about people logging in — who signs in, which screens they see. An API key governs machine-to-machine access; it works in the background, with nobody watching a screen. Both live under the heading of "access," but they guard different doors. This piece is about the second door.

Picture a hotel key card: the front desk hands you a card, it opens only your room, and it is cancelled at checkout. A well-managed API key should behave the same way — limited, traceable, and revocable when its time is up. The problem is that most small businesses mint that card once and then forget it forever.

Where leaks actually come from

In a small business, a leaked key almost never looks like a heist movie. It usually happens in far more ordinary ways:

  • The key pasted into a chat: WhatsApp, Telegram, email — it doesn't matter. Once sent, the key sits there indefinitely, and anyone with access to that thread can read it.
  • The screenshot: You photograph a settings screen and drop it into a shared drive; the key is quietly sitting inside that image.
  • The departed employee or freelancer: The work ends, but the access doesn't. A key nobody revoked stays live for months, sometimes years.
  • The single "master key": You use the same key for every integration, so when one leaks, all of them are exposed at once.

The common thread: the problem is rarely an outside attack and almost always lost control on the inside. Old keys are the sneakiest, because they keep working silently while nobody notices. The good news is that every one of these is fixed by a few simple habits, no expensive security software required.

Give each key only what it needs

The simplest rule in security: give a key the least it needs to do its job. This is called least privilege, and while the name sounds technical, the logic is plain.

Say you create a key so your website form can add new contacts to your CRM. That key needs one permission: "add a contact." It has no business deleting records, viewing invoices, or managing users. These permissions are called scopes, and most decent tools let you tick them one by one when you generate a key.

Where you can, prefer read-only keys. If a reporting dashboard only needs to read your data, don't hand it write access. Keeping the scope narrow means that even if a key leaks, the damage is trapped in one small room instead of the whole building.

A separate key for every integration

One "opens-everything" key feels convenient, and that is exactly why it's dangerous. When it leaks, every connection you have is exposed at once, and you can't tell which system was at fault.

Instead, give each connection its own key: one for your online store, one for accounting, one for your website form. Then, when you revoke a key, only that one connection stops — the rest of the business keeps running.

CRMOnline storeAccountingWeb formPhone
Each connection speaks with its own key, so one leak doesn't touch the others.

This structure also helps when you decide whether to build a connection natively or with a third-party tool: treating each connection separately lets you manage its risk separately, too.

The life of a key: create, use, rotate, revoke

A healthy API key has a life cycle. The key you generate once and forget becomes, over time, your biggest exposure.

1Create2Limit scope3Store safely4Rotate regularly5Revoke
A healthy key moves through these five stages.

Rotation means retiring the current key and replacing it with a fresh one. Do it in two situations: on a regular schedule (pick a rhythm you can actually keep — quarterly is a sensible starting point for most small teams) and after an event (a staff member leaves, a device goes missing, or you suspect a leak — rotate immediately).

The trick to painless rotation is order: generate the new key, switch the integration over to it, confirm it works, then revoke the old one. Done in that sequence, no connection ever drops.

Storing keys safely

You've generated the key and narrowed its scope — so where do you keep it? The wrong answers: in email, a chat, a spreadsheet, or a "keys.txt" on someone's desktop.

The right approach takes a little discipline but isn't complicated:

  • Use a password manager: A password manager with a shared team vault is the most practical home for keys, and it shows you who has access to what.
  • Leave the key inside the tool itself: Most CRMs and automation tools store the key in their own secure area; you rarely need to copy it out at all.
  • Keep it out of chat and email: If you must share a key, use a secure, non-permanent method — never paste it into a group thread.

By the way, automation platforms like Zapier and Make also hold the keys of the services you connect to them. That's convenient, but it means something: whoever can access that platform effectively has access to those keys. Keep the access list clean.

Manage access from one panel

Rocketly lets you see your integration connections and their keys, with their scopes, in one place.

Explore Rocketly

Who can get in?

A question matters as much as the keys themselves: who can create, see, and revoke them? In small teams it feels natural for everyone to have access to everything, but integration keys should be the exception.

Concentrate the power to create and revoke keys in one or two trusted people. The rest of the team can use the integrations without ever seeing the keys. That way, when someone leaves, there's a single place to check.

And the step most often skipped: offboarding. When an employee or contractor leaves, rotate the keys they created or knew about. When the person who set up your accounting integration moves on, that connection's key should be refreshed on the way out.

If a key leaks: the first five minutes

One day you may notice a key has been exposed — in a screenshot, an old email, or as an unexpected bit of activity. Don't panic; move quickly and in order.

  1. Revoke the key immediately. Disabling a leaked key is always safer than "keeping an eye on it."
  2. Generate a new key and restore the connection. Remember to keep its scope narrow again.
  3. Check the logs. Look for anything unusual done with that key.

Fast recovery has a quiet prerequisite: regular backups. If a problem touched your data, a solid backup and disaster recovery plan is what protects you from the real loss. Security isn't one measure; it's small measures stacked on top of each other.

A quick checklist for small businesses

If you do nothing else, run through this short list once and you'll be ahead of most small teams:

  • One key per connection, never a single master key shared across every integration.
  • Narrow scopes, read-only wherever a tool only needs to read.
  • A rotation rhythm you can keep, plus an immediate rotation after anyone leaves.
  • Keys stored in a password manager or the tool itself — never in chat, email, or a screenshot.
  • A short list of people allowed to create and revoke keys.

None of this requires a security team or a big budget — five habits that, once in place, mostly run themselves.

Frequently asked questions

Is an API key the same as a password?

Similar, but not the same. A password is for a human to log in; an API key lets two pieces of software recognize each other. Both must be kept secret, but because no person types a key in each time, rotating them and narrowing their scope matters even more.

How often should I change my keys?

Pick a rhythm you can keep; for many small teams, once a quarter is a sensible starting point. Beyond that, rotate without waiting whenever someone leaves, a device goes missing, or you suspect a leak.

What is a read-only key for?

A read-only key can read data but not change it. It's ideal for tools that only display information, like reporting or dashboards; even if it leaks, no one can delete or corrupt your data with it.

Does a small business really need all of this?

Honestly, if you have two integrations, you don't need an enterprise-grade secrets manager. But three simple habits — separate keys, narrow scopes, and regular rotation — are worth it for everyone and take almost no time.

Managing API keys really comes down to putting a few small habits in order: a separate key for every connection, only as much power as each key needs, and a clean rotation now and then. Once those are in place, growing your integrations won't cost you any sleep. If you run a CRM like Rocketly, seeing your connections and access rights in a single panel makes those habits easier to keep — because you can't manage a key you can't see.