Sessions
A working conversation with something that already knows your context. Ask in plain language, watch it read the signals, and leave with the work done instead of a summary of what you should do.
Talk it through, leave with it done
A generic chatbot starts from nothing every time. You paste in the plan, explain who the athlete is, describe the last four weeks, and then get advice that reads like it came off the internet, because it did. Then you go and do the work yourself anyway.
A session starts from what it already knows: your history, the data flowing in from the apps you use, and the plans sitting in your workspace. So the question is short and the answer is specific — and the plan is updated by the time the answer lands.
We give the read. You make the call. The point isn't to take the decision off you, it's to make sure you're making it with everything in front of you.
What a session is for
The same room handles the programming, the triage, and the admin around both.
Write the week, then change it
Ask for the block, argue with what comes back, and adjust it out loud. COORDINATOR writes into the plan as you go, so you finish the session with the document updated rather than a transcript you still have to act on.
Ask what the data is telling you
Someone's recovery has been red for four days, someone else's paces are drifting late in the run. Ask what to do about tomorrow and get a straight read on the signals, with what it looked at shown above the answer.
Clear the work around the work
Summarise a week of check-ins, turn an intake form into a starting block, draft the message you've been putting off, flag who's renewing. The part of the job that eats evenings.
Go back over what happened
Pull up a past session and keep going from it. Sessions are kept, named, and searchable, so the thinking behind a decision six weeks ago is still there when someone asks why.
Sessions inside a group
In a group, a session happens in a shared room with everyone's COORDINATOR present. Post the week once, let members ask their own questions against it, and share the findings worth sharing.
It doesn't have to be training
A session is a working conversation with something that knows your context. People use it for programming, but also for planning a season, prepping a consult, pricing a package, and thinking out loud on a drive home.
Every turn shows its work
Before a single token is generated, the turn gets appraised: what you're asking, how much reasoning it deserves, how deep to pull from memory, and which model is built to answer it. You can override that yourself — Quick for speed, Deep for the hard ones, Auto to let it choose.
Then it tells you what it did. Which model answered, what it read, which tools it ran, and what it wrote. That timeline sits above the answer, not buried in a settings page, because you can't trust a call you can't audit.
Attach anything the question needs: a file off your device, something from your library, an artifact from your workspace. When a session produces something worth keeping, it arrives as a draft and stays one until you save it.
Your sessions are yours
A personal session is private to you. It stays that way in a group: your COORDINATOR runs on your own history, and what reaches the room is what you decide to share there.
Pin what you're living in, rename what you'll need later, and delete anything you don't want kept.