Connector categories

The constraints that apply to a whole category, not one vendor.

Most of what shapes an integration is a category property: a CRM has ownership and duplicate semantics a chat tool does not; a ledger has period-close rules a support desk does not. Designing against those properties is what makes an integration survive the vendor being replaced.

Category-level

Constraints that hold across vendors

Design-first

What to decide before choosing a platform

Vendor-neutral

Applies whichever product you run

Introduction

Why category is the more useful unit.

Teams usually start from a vendor name, but the decisions that determine whether an integration survives are category decisions: what the system is authoritative for, what it emits that is worth acting on, and which failure mode it produces under load. The vendor determines the API shape. The category determines the design.

Getting those right transfers if you later change vendor within the category — the workflow gets repointed rather than rebuilt. Getting them wrong is expensive regardless of which product you picked, because the cost lands in the workflow rather than in the connector.

The useful test is to describe your integration without naming the product. If the description still makes sense — this system owns the customer record, it emits an event when the stage changes, it accepts a write on the activity object — you have designed against the category. If it collapses into a list of endpoint names, you have designed against a vendor, and the next migration will start from scratch.

That is also why these pages sit above the per-platform guides rather than beside them. A category page is meant to be read once per category and then reused across every integration you build in it; a platform guide is meant to be read once per implementation.

Unit
System category
Applies to
Any vendor in the category
Deeper
Per-platform guides
Narrower
Live connector pages

Why this exists

Vendor-first integration design does not survive a vendor change.

An integration designed around one product’s quirks encodes those quirks everywhere. When the product is replaced — and in most of these categories it eventually is — the workflow has to be rebuilt rather than repointed, and the rebuild is charged to a team that has already moved on to something else.

The second failure is authority drift. Nobody decides which system is authoritative for a shared field, so two of them write it, and the integration becomes a race whose winner varies by sync timing. This is invisible until the day someone reconciles the two and cannot explain the difference.

The third is designing for the happy path of a category whose defining property is a specific failure mode. Ledgers have period close. Support desks have ticket merges. Chat platforms have rate limits and message ordering that is not guaranteed. Each of these breaks a naive integration in a way that reads as a bug in the workflow rather than a property of the category.

Designing against the category constraint instead keeps the vendor-specific part at the edge, which is both cheaper to change and easier to reason about when something goes wrong at two in the morning.

You're likely here because

  • A vendor change is plausible within a few years
  • The same integration logic is being rewritten per product
  • Nobody has decided what the category is authoritative for
  • Two systems both write the same field on the same record

How to choose

Where to go from here.

These pages are ordered by how much you have already decided, not by category size.

If

You know which systems participate but not how they should behave

Start with

Pick the category below and read what it owns, emits, and accepts

If

You have chosen the platform and need its specifics

If

You want connections we can vouch for end to end

If

You are still deciding what the workflow should do

If

You need the decisions written down before anyone builds

How it works

What every category page answers.

The same five questions, in the same order, for every category. The order is deliberate: you cannot decide which writes are safe before you have decided what the system is authoritative for.

01What it owns02What it emits03What it accepts04How it fails05What to measure

Step 01

What it owns

The records the category is normally authoritative for, and the fields that other systems should therefore treat as read-only. Getting this wrong is the source of most reconciliation work later.

Step 02

What it emits

The events worth triggering work from, how delivery behaves under load, and whether ordering is guaranteed. A category that emits at-least-once needs idempotent handling downstream.

Step 03

What it accepts

Which writes are safe to automate, which need an approval boundary, and which should never be automated because the category has no way to undo them.

Step 04

How it fails

The category-specific failure mode that shows up regardless of vendor — period close, merge semantics, rate limits, partial sync — and what a workflow should do when it hits one.

Step 05

What to measure

What completion looks like for workflows built on this category, so a run that half-succeeded is visible as a failure rather than reported as a success.

Scope

Limits of category-level guidance.

  • Category constraints do not cover vendor-specific API behaviour, quotas, or authentication quirks.
  • Availability still depends on your provider configuration and the scopes you granted.
  • A category page is a design aid, not an implementation guide — it will not get you to a working connection on its own.
  • Where a vendor deliberately breaks the category pattern, the per-platform guide is authoritative and the category page is wrong for that product.
  • Nothing here replaces testing against your own data. Category constraints predict the shape of the problem, not its size in your environment.
  • These pages describe systems as they behave today; a category’s constraints move when the dominant products in it change their model.

FAQ

Questions about this collection.

Why not just read the vendor guide?

Read both. The category page tells you what to decide; the vendor guide tells you how that platform behaves. Skipping the first is how integrations get rebuilt after a vendor change rather than repointed.

How many categories are there?

Eight, chosen because they have genuinely different authority models and failure modes rather than to fill a grid. A ninth that behaved like one of the existing eight would be a page that made the set look bigger and the guidance no better.

My system spans two categories.

Common — many platforms do. Read both, and decide per record which category’s rules apply to it. The mistake is deciding at the system level, because that forces one set of rules onto records that do not follow them.

Does this replace provider documentation?

No. It is design guidance; the provider remains authoritative on API specifics, limits, and behaviour. Where the two disagree about a specific product, the provider is right.

Do we need UbiVibe for any of this?

No. The category constraints apply to any implementation, in any stack. They describe how these systems behave, not how our runtime behaves.

Where should design actually start?

With the workflow. The category matters once you know which systems have to participate — starting from the connector list produces an integration in search of a job.

Start with ARIA

Ask ARIA about connector categories.

You do not have to pick your way through this collection to get started. Describe the outcome you want and ARIA determines which capabilities, systems, and workflows the job needs.

  • ARIA acts only through the systems and permissions you connect.
  • Connections use scoped credentials you can change or revoke.
  • Actions are recorded, and consequential ones can require approval.

Goes to UbiGrowth, with the page you asked from attached. We do not sell or share it. Prefer to talk? Call 972-823-1294.

Start here

Design against the category, not the logo.

Describe the handoff you need. ARIA resolves which categories have to participate and which of their constraints will shape the design.