What we know
Tools we've used properly, not just watched a demo of.
Anyone can list logos. Here's the same list with an honest opinion attached to each one, because knowing when not to use something is the useful part.
Automation
APIs
Sooner or later the ready-made integration doesn't do what you need. That's when we go to the API, read the documentation properly, and build the connection that was actually described in the meeting.
What it's good at
- Doing exactly what you asked for, rather than nearly
- Access to fields and actions the no-code connectors never expose
- Better performance and control at volume
- Being explicit about errors, limits and retries
Where it gets awkward
- Rate limits and pagination catch people out at scale
- Authentication and token refresh add moving parts
- Providers change things, so integrations need looking after
- Documentation is often optimistic about how it really behaves
Our experience with it
We've built against plenty of well documented APIs and a few that clearly hadn't been read by their own authors. Either way, we build them to fail loudly and recover cleanly.
Things we've built or connected
Direct, server-side integrations that keep credentials out of the browser
Custom syncs where the standard connector didn't go far enough
Connections built to handle rate limits without losing data
Usually in the same conversation
