Newimproved

Choose how you want to talk: nine modes, from Listen to Boundaries. Practice first with an AI. What stays, in three shelves.

Every talk now starts with a goal and a mode: Understand, Work it out or Resolve, then one of nine modes from Listen to Boundaries. Before a real talk you can practice a difficult situation with an AI partner and get feedback at the end. What stays from a talk now lands in three shelves, Agreements, Insights and Relationship, and each of your views carries one honest word: held up, changed, let go or still open.

Start a talk

The door.
Connect opens with one card: describe your situation, pick a mode, tap Structure. The mode picker asks for your goal first (Understand · Work it out · Resolve), then the mode: Listen, Dialogue, Be heard, Debate, Decision, Negotiation, Settle, Repair, Boundaries. Any mode keeps it open until the next step.
270 new topics
in English, German and Chinese, filtered by life area (Family, Work, Home, Politics, Values) and by your mode. Cards show how many people are waiting right now, and Read more opens the background.
For people who know each other.
Negotiation, Settle and Boundaries are not matched with strangers: the card says Invite someone and hands the topic to your invitation.
A gauge that fits the topic.
Agreement with faces, two sides with planets, a number (a salary, a budget) with a free-moving needle, or a temperature that stays with you alone. Listen and Be heard ask for a role instead: I'll share or I'll listen.
Requests out of the way.
Your waiting requests no longer sit in the talks list. A pill in Connect, "n requests open · view", shows them with a way to withdraw.
Joining by code
shows a preview first (title, person, mode) and says clearly when a talk cannot be joined.

Matches and talks

Looking for a match
shows two drifting planets.
You have a match!
shows photos, your two positions as pills and three buttons: Prepare, Start now, Other topics.
Match circles.
A match that has not started yet sits as a circle above your talks. Running talks are on top, ended talks below.
One way of waiting.
Loading dots, skeleton rows and, while the AI answers, three dots in a bubble with the AI badge. Sending is locked meanwhile, so nothing goes out twice.
The mode sits as a chip in the talk header. Tap the name to see your partner's profile card.
Breathe for a moment.
In preparation, three guided breaths replace the mood question, open by default in the Resolve modes. In the Settle tool you can ask the mediator for a wink of humour.
The mediator's first message names the mode and, where it matters, the two sides or the unit. Pictures and videos in talks are gone.

Reflection and results

An AI-drafted agreement.
The reflection starts with a proposal of the right type (Agreement, Boundary, Reconciliation, Acknowledgement, Postponement) in three editable lines: your position, their position, what applies now. Send it to your partner or choose No agreement. Insights come as chips: for me, or propose together.
Has your view shifted?
Answer on the gauge (Held up or Changed) or with two buttons: Let go and Still open.
And between the two of you?
Closer, Unchanged, Strained. Only you see your answer.
Three shelves.
Results opens with three tiles: Agreements (need both of you), Insights (yours, or confirmed together) and Relationship. Cards look like dated documents, proposals wait for confirmation, and the only number is "n waiting for your confirmation". After a reflection you land right here with the chip Talk with {name}.
My Views
gives each position one word, Held up, Changed (with how far), Let go or Still open, and a small line saying which talk it came from.

Practice with an AI

AI practice talks.
A new area in Practice with three scenarios: Asking for a raise, Politics at the family dinner, A fight with your partner. Five to ten minutes, up to twelve exchanges, the AI badge on every message, and a feedback card at the end: what went well, what to try next.
One AI practice talk per month is free. With Pro you get up to 30 talks + 30 AI practice talks per month, at most 3 talks and 5 practice talks per day.
The transcript is deleted 30 days after the end, your feedback stays. A very short practice talk is not counted.

Safety, rights and your account

Report any message
with nine clear reasons. Under Settings you see the status of your reports: received, result, and how to appeal. AI suggestions and the mediator can be reported too.
If a message is hidden or a talk is ended, you see the rule, the time and the way to appeal. A suspended account can still read its talks and delete itself.
Sign-up explains
what WorldViews does with sensitive data and with the AI (Google Gemini) and asks for your consent, the rules and 18+. Existing accounts confirm once. A short notice says that an AI takes part.
My profile is a showcase.
A sentence about you, up to three talk-style chips and badges, each shared only when you switch it on. Others see only what you share.
Settings in seven groups.
Legal now lists Terms of Use, Privacy Policy, Community Rules, Imprint and Child Safety Standards, plus a contact line. New: Request your data, open-source licences, a support-access switch, a notice before the Canny feedback board, and on the web the forms to cancel or withdraw a contract.
Less data by default.
Crash reports are opt-in and asked once on your device. No usage analytics, no microphone permission, fonts shipped with the app.
On the web, WorldViews is currently available in Germany, Austria, Switzerland and the USA; the country is read from your connection and never stored.

New here?

Sign-up now begins with a welcome and three short screens, then You (name, photo, a sentence), then Where to start: try a practice talk, invite someone or pick a topic. Each place in the app explains itself once.

Languages and design

Everything new comes in English, German and Chinese. Chinese wording is unified (你, 账户).
A new look across the app: cards, chips, pill buttons, readable dark mode, a visible back arrow on the web, and the mediator's planets wherever two sides meet.