a reflective chat that brings clients to your session with more to start from.
Blueno is a chatbot built not to flatter the user. People write in their own words, in their own time, and the system holds what they put down without advising, scripting, or scoring them. It is built for the in between, the things that come up between sessions, the half-formed thoughts that do not survive a wait until next week.
Theodore Addo
Brown's Warren Alpert Medical School '26. Incoming PGY-1 at MetroWest Medical Center, matched into diagnostic radiology at St. Vincent Hospital.
Shadrack Annor
Brown CS + Religious Studies '27. Researches AI sycophancy at SWRL, Brown's sociotechnical lab.
What blueno is
The product itself is live at blueno.polarity-lab.com as a free research preview. It is not a clinical tool. It does not deliver a protocol. It does not produce a diagnosis. We are careful about that line on every surface a user touches.
What it does, instead, is treat reflection itself as the material. Blueno runs on Polarity Lab's Cosmos engine — the same account-level graph as Cosmos, GPS, WaxFeed, and AVDP — and turns ongoing conversation into short passages, written from the client's own words, that they can read back later. The unit of memory is not the message. It is the passage.
What the clinician layer is
A reading surface, not a dashboard.
We are building a small, read-only view that lets a clinician see what their client has chosen to share. The unit of sharing is the Chronicle, a short passage Blueno writes between sessions from accumulated material the client has produced (their captures, their chat, both). A Chronicle is never shared automatically. Your client picks which ones reach you, opts in per passage, and can revoke any time.
We are not building a charting system, a billing integration, or anything that would put Blueno inside the medical record. The point of contact is the conversation you already have with your client. Blueno just gives that conversation more to start from.
The Chronicle itself, a short passage of Blueno's writing (typically a paragraph, 200 to 300 characters) drawn from your client's material. Up to eight source excerpts attached, each a verbatim quote from a capture or chat message your client wrote, timestamped and tagged by source, so you can see exactly what the model read. An optional reflective question Blueno generated alongside the Chronicle. An optional note from your client when they share.
Behavioral metrics. Engagement scores. Sentiment graphs. Diagnostic suggestions. Streak counters. Anything from your client's chat history or captures that they have not explicitly attached to a shared Chronicle.
Consent runs at the Chronicle, not the account. Per passage, opt-in only, revocable at any time. There is no account-level setting that turns on a steady pipe to you in the background.
No DM channel between clinician and product. No nudges sent to you. No notifications. No charting or billing integration. The view is a reading surface, not a workflow.
An example
What a Chronicle looks like.
Synthesized for demonstration. Not from any real client. The shape, length, and source-citation pattern match what the backend actually generates.
you keep noticing the same pattern with your sister, the way a small comment turns into a longer silence. you say it does not bother you and then you find yourself replaying it later. there is a difference between letting something pass and pretending it did.
“talked to my sister yesterday and she said something that felt off but i did not push it. now i am wondering why i never push it.”
“i used to think being the calm one was a strength. lately it feels more like a habit i do not know how to break.”
what would it look like to push, just once, and see what happens after?
Wanted to bring this in for our next session. The line about the habit is sticking with me.
Who we are looking for
A small number of design partners for v1.
Therapists who already think with their clients about what happens between sessions. Integrative and somatic practices. Digital health partners curious about AI as something quieter than a chatbot. We are listening more than selling.
Reach out either way below. A real person on the Polarity Lab team will reply.