Loading...
Human-first AI reflection
Most companion AI is rewarded when a conversation never ends. Yatri AI is designed for the opposite outcome: name what is happening, find one grounded next step, and deepen connection with real people rather than replacing them.
Direct answer
Yatri AI is Soul Yatri’s reflection companion. It can help a person put feelings into words, prepare for a conversation, notice a recurring pattern, or decide which human support path fits. It is not a therapist, doctor, emergency service or human friend. It does not diagnose; separate insight-save and side-effect tool actions require an explicit confirmation flow.
What the product actually does
01
Yatri AI is built to help the user understand a situation and take a grounded action outside the chat. Its base safety boundary defines it as a reflection companion rather than a therapist, emergency service or human friend, and instructs it not to pressure the user.
02
When a person asks Yatri to create a saved insight, the product proposes that tool action and waits for explicit confirmation. The user can reject it. Ordinary account and conversation data remains subject to the applicable privacy policy and product controls.
03
Messages are checked for crisis and other high-risk categories. Crisis language takes a deterministic safety route rather than relying on an open-ended model response. The purpose is immediate signposting and escalation, not automated clinical judgment.
04
A useful reflection can lead to a journal prompt, a clearer message, therapy discovery, a Soul Circle request, or a relationship check-in. Yatri AI is the connective layer; the actual life change still happens through human decisions and human relationships.
How it works
Describe the conversation you are avoiding, the feeling you cannot name, or the decision that keeps looping. Yatri works best with concrete context rather than a demand for a diagnosis, prophecy or absolute answer.
The response should separate facts, feelings, assumptions and possible next actions. The aim is not to win an argument for the user. It is to make the situation legible enough that the user can choose what to do.
Side-effect tool actions are proposed, not silently executed. Creating a separate saved insight requires confirmation. A drafted message still belongs to the user to edit and send. This pause is a product feature because emotional software should not act faster than the person can think.
Safety and product boundary
Yatri AI is deliberately narrower than clinical or emergency support. It cannot provide a diagnosis, prescribe treatment, support dating or sexual flows for minors or people whose adult status is unknown, or handle an emergency as if it were a trained responder. High-risk content is routed to a safety response and report path. The product can help a person find words and options; it cannot make a human judgment on the person’s behalf.
Where it can help
Sort what happened from what you fear happened, identify the request you want to make, and draft language that is direct without being needlessly escalatory.
Name the feeling, notice the repeated interpretation, and choose one action that changes the next ten minutes instead of asking the chat to solve an entire life at once.
Decide whether the next step is a friend, a partner, a therapist, a community match, a journal entry, or urgent local support. Yatri can help frame the decision but should not become the destination.
Common questions