Agentic

Just ask your agent to run the session

You do not wire anything up. Point your agent at pointstack.dev and tell it what you want in your own words. It reads /llms.txt the way it reads any docs, and takes it from there. Works with Claude, ChatGPT, Cursor, or anything else that can browse and call an API.

Things to say

You:Put these five tickets up for estimation on pointstack.dev and send me the room link. Check the site’s llms.txt for how to.
The first ask is the only one that has to say where to go. It reads the reference, creates the room, adds the stories with their ticket links, and hands back a URL you can paste into chat.
You:Add the two bugs I just filed to that estimation room.
No need to repeat yourself: it kept the room and the token it was given, so a follow-up needs neither the site nor a room code.
You:What did the team land on for the billing work?
It reads the room back and tells you the estimates, once people have voted.

Agents cannot cast estimates. Nothing in the API sets a story's points. Those come from people voting in the room. An agent runs the logistics; the judgement stays with your team.

What your agent needs from you

Usually nothing. Creating a room takes no key at all, which is why "put these tickets up" works the first time you ask. Two kinds of token exist for when you want more than that.

Room token

automatic

Handed to your agent when it creates a room, and good for that room only. It is how "add two more tickets to that room" works an hour later. You never see it unless you want to: a moderator can revoke it in room settings.

Account token

you generate

Lets an agent act as you across every room you own, and create rooms in your teams that do not expire. Generate one on your Account page and paste it to your agent once. Rotate or revoke it there whenever you like.

What you get

Nothing to install
Your agent reads /llms.txt and knows the whole API: what to call, what comes back, and how to recover from an error. There is no plugin, SDK or setup step.
Room rules, set on creation
Deck, auto-reveal, auto-finalise, auto-play, guests, spectators, and how many people may moderate or vote. Ask for a room where everyone votes and it will set that.
Your team just votes
An agent cannot cast an estimate. It opens the room, fills the backlog and reads results; the numbers come from people in the room.

If you want the details

Everything an agent needs is written for it to read start to finish: /llms.txt for the API, workflows.txt for common jobs end to end, and errors.txt for what to do when something fails. You are welcome to read them too.