#490 – Introducing the Parabol API

This week we launched Parabol’s API into general availability.

Eleven years is a long time to build a piece of software before opening it up to other developers. Probably longer than it needed to be. But we’ve been deliberate about this: we wanted the API to reflect something stable and real, not a moving target. We think we’re there.

Starting today, any Parabol user—free or paid—can generate a personal access token directly from their profile settings. The API is GraphQL-based and interactive; you can explore it live at action.parabol.co/graphql. We built it with AI-agent-friendly metadata, which means agentic workflows can introspect what’s available without heavy documentation-reading. Paid accounts get higher request rate limits. Free accounts get access too.

To show what’s actually possible, we built a reference app called retro-reflect-bot. It takes unstructured text—Slack history, knowledge base content, customer service logs—combines it with GitHub activity (feel free to drop us a PR for other services!), and generates retrospective reflections from it. The problem it solves is real: retrospectives are only as good as what people can recall, and human memory is biased toward the recent and the dramatic. retro-reflect-bot shows how pulling in ambient data can help a team reflect more honestly on the full shape of a sprint.

Here’s a demo video:

Beyond retrospectives, the use cases we’re already hearing about include exporting standup data in bulk, syncing Parabol to and from proprietary internal systems, and handling complex enterprise provisioning that goes beyond what SCIM supports today. If you can think of a need, there’s likely a path to build toward it.

OAuth 2.0 is next on the roadmap, which will extend the authentication pathways available to AI agents. If you’re building something on the API, or just want to talk through what might be possible, reach out at parabol.co/contact. We’d love to hear what you’re working on!

A sharp decline in meetings completed this week; such a sharp decline in fact we’re wondering if there is a bug in our metrics gathering pipeline…

…shipped the Parabol API into general availability. Personal access tokens are live in profile settings.

…began planning our next development cycle. We ended up ahead of schedule this development sprint and are advancing our next planning cycle to take on more ambitious work.

…upgraded our development infrastructure. When gathering the metrics for this week, we noticed that some of our data marts had gone stale. It turned out one of the nodes in our Kubernetes clusters had ran out of disk space, and our monitoring infrastructure didn’t catch it.

…further automated operations. We have an internal bot that keeps an eye on things like accounts receivable and unaddressed customer success interactions. We enhanced this bots capabilities this week to be of further help to our teams.

…prototype out some of the bets we’re interested in investing in for the next development cycle.