A specific message contains the useful part.
A friend mentions she is waiting for a Windows auto-update result. The whole chat does not need to become memory; only the reusable care point does.
Private AI context for me and us
Private context that syncs across devices and still works offline.
Soulife keeps relationship context, notes, and AI actions close to the devices where your life happens: encrypted by default, available across Mac, Windows, iOS, and Web, and useful even when the network is not.
Developer preview downloads Mac · Windows · iOS · Android · Web Encrypted by default Private conversations and source references are designed around E2EE instead of server-readable history.
Same context everywhere A care point saved on desktop can come back on iPhone, Web, or Windows when the relationship continues.
Still useful offline Local context remains readable when the network drops, then syncs confirmed changes later.
Private by design End-to-end encrypted conversations and source-visible AI context.
Multi-device sync Continue relationship context across Mac, Windows, iOS, and Web.
Offline usable Local context remains readable and actionable, then syncs later.
Worth a natural follow-up AI found 2 source clues
She is waiting on the Windows auto-update result. Do not push; ask softly.
Soulife turns the useful part of a chat into a care point, keeps the source visible, and helps you bring it back at the right time.
Source: May 15 chat mentioned update stability.Rhythm: no follow-up for a week; low-pressure check-in fits.Boundary: draft only, never auto-sent.
Save to Maya · Recent care point
Save care point View source
One complete loop
From a tiny chat signal to a natural next line.
The page should not ask users to imagine a database. It should show the moment they recognize: someone said something small, it matters later, and Soulife remembers it without turning the relationship into a task board.
Next natural line “Did the Windows auto-update settle down after last week?”
No auto-send. No hidden server-side reading. The source remains visible.
Soulife asks how to remember it.
Before writing anything, AI asks whether this should be a care point, reminder, boundary, or relationship direction.
The user sees where it will be saved.
The preview shows the exact target, source message, and wording. Nothing is written until the user confirms.
It comes back as one natural sentence.
Later, the relationship radar suggests a low-pressure line and explains why this is the right rhythm.
Why people come back
Show the moment users can feel, not a feature matrix.
Soulife is easiest to understand through small moments: a friend sounds uncertain, a parent mentions a health detail, a shared household preference disappears into chat. The product turns those moments into private, reusable context.
Maya is trying not to overthink a career change.
Ryan keeps a close but low-pressure friendship with Maya. They usually talk by voice notes and short messages.
Source
Maya: “I think I want to leave, but I also feel guilty because the team is already stretched.”
Next moment Next week, Relationship Radar suggests: “Lightly ask how the decision is feeling now. Avoid solving it unless she asks.”
A parent mentions a small health detail that is easy to forget.
Nina lives in another city and wants to care without turning family conversations into a checklist.
Source
Dad: “The new medication makes me a bit dizzy in the morning, but it is probably fine.”
Next moment When Nina opens the chat later, Soulife surfaces a natural line: “How have mornings felt since the medication change?”
A practical message becomes shared context instead of disappearing in chat.
Alex and Sam use one shared space for household context, but private notes stay private on each person’s device.
Source
Sam sends a photo of a receipt and says: “This is the cleaner we liked. It did not trigger allergies.”
Next moment Two months later, AI can answer “what cleaner did we like?” with the source and the person who confirmed it.
Interaction frequency becomes a human rhythm, not a score.
Sofia and her older cousin talk a few times a month. The cousin recently mentioned visa paperwork stress.
Source
Signals: 3 chats in 2 months, 2 started by the cousin, and an open topic about paperwork due this week.
Next moment Relationship Radar shows a single suggestion with the reason: ask once, then stop unless she expands.
AI drafts in the user’s communication style, but never sends.
Martin is writing to a mentor. He wants to sound grateful, specific, and concise without asking for too much.
Source
Style context: start with thanks, add one concrete update, and avoid a message that requires an immediate reply.
Next moment AI produces a restrained draft that sounds like Martin, and the user can edit or ignore it before sending manually.
Not a CRM
It should feel like caring better, not managing people.
Soulife avoids scores, queues, and automatic outreach. It uses context to suggest a natural rhythm, keeps drafts under user control, and lets some relationships stay quiet.
Natural rhythm instead of scores
The product says “lightly ask once this week” or “give it space,” not “relationship temperature 62/100.”
Drafts instead of auto-send
AI can help phrase a message, but the user edits, copies, or ignores it. Nothing leaves the app automatically.
Source-visible context instead of surveillance
Every reusable memory can show why it exists: source message, note, imported chat fragment, or confirmed user edit.
Product proof
The promise is visible inside the app, not just on this page.
Soulife’s strongest loop has to be testable: find a source, ask how the user wants to remember it, preview the exact write, confirm it, then read it back from the relationship or space later.
Source stays attached
The useful context keeps a link back to the chat, note, imported fragment, or shared space that created it.
AI asks before writing
Instead of turning every signal into a hard task, AI first asks whether it should be remembered as a care point, boundary, goal, reminder, or draft style.
The user confirms the write
The app shows the target location and wording before anything becomes reusable context. Saved records can be opened from the relationship detail later.
Context comes back later
The next suggestion can explain why it appears, what source it used, and why the rhythm is gentle, quiet, or ready for a deeper conversation.
Runnable fixture 10 loops verified in product
Source message
“The Windows auto-update still feels unstable. I am waiting to see if it settles before changing anything.”
AI asks one narrowing question
“Should I remember this as a care point, reminder, relationship direction, or communication boundary?”
Preview before saving
Save to Maya · Recent care point. Reason: it gives a natural future follow-up without pushing.
Read back later
Relationship Radar suggests one soft line this week and shows the source chain behind it.
Writes only after Preview + confirm
Readable on Mac · Windows · iOS · Web
Trace includes Source · prompt · action
Three commitments
Concrete scenarios only matter if the foundation is private, synced, and resilient offline.
01
Private by design
Privacy protects the sensitive parts
Relationship memory, imported chats, private notes, and AI prompts need a privacy model first. Soulife keeps sources visible and makes AI write actions explicit.
4 verified loops
02
Multi-device sync
Sync keeps context with you
A care point saved on desktop should be available when you reply from your phone. The product is designed around personal and shared context moving across devices.
3 verified loops
03
Offline usable
Offline keeps the product dependable
Important context should not disappear in a weak network moment. Local-first reading and later sync make the product useful during travel, commute, and everyday interruptions.
3 verified loops
Concrete loops
The product starts to matter when one small piece of context comes back at the right moment.
These are simulated examples, but they mirror the loops Soulife is built around: a chat, a note, or a shared space becomes context that can be confirmed, reused, and traced back to the source.
Private by design
Privacy protects the sensitive parts
End-to-end encrypted conversations and source-visible AI context.
01 Close friend
Maya is trying not to overthink a career change.
Ryan keeps a close but low-pressure friendship with Maya. They usually talk by voice notes and short messages.
Source
Maya: “I think I want to leave, but I also feel guilty because the team is already stretched.”
AI asks
Soulife asks Ryan: “Do you want me to remember this as a care point, a boundary, or a decision in progress?”
Saved context
Saved to: Maya / Relationship space / Recent care point. “Maya is weighing a career change and may need low-pressure emotional support, not immediate advice.”
Next moment
Next week, Relationship Radar suggests: “Lightly ask how the decision is feeling now. Avoid solving it unless she asks.”
02 Family
A parent mentions a small health detail that is easy to forget.
Nina lives in another city and wants to care without turning family conversations into a checklist.
Source
Dad: “The new medication makes me a bit dizzy in the morning, but it is probably fine.”
AI asks
Soulife detects a sensitive follow-up and previews a softer memory instead of a hard task.
Saved context
Saved to: Dad / Care context. “Dad mentioned morning dizziness after changing medication. Ask gently next time whether it improved.”
Next moment
When Nina opens the chat later, Soulife surfaces a natural line: “How have mornings felt since the medication change?”
04 Personal
A private note becomes reusable self-context.
Jordan is trying to understand energy patterns without sending that information to a shared relationship space.
Source
Private note: “When I start the day by checking messages, I feel reactive for the next few hours.”
AI asks
Soulife asks whether this should become a routine insight, a warning sign, or a question to revisit.
Saved context
Saved to: Me / Behavior routine. “Morning message checking makes the day feel reactive; try writing first before opening chat.”
Next moment
The AI assistant can later connect this with local activity context and suggest a gentler morning rhythm.
08 AI persona
AI drafts in the user’s communication style, but never sends.
Martin is writing to a mentor. He wants to sound grateful, specific, and concise without asking for too much.
Source
Style context: start with thanks, add one concrete update, and avoid a message that requires an immediate reply.
AI asks
Soulife asks: “Should the draft lean grateful, direct with an update, or exploratory with a question?”
Saved context
Saved to: AI persona draft. “Mentor drafts should be short, grateful, concrete, and copy/edit only.”
Next moment
AI produces a restrained draft that sounds like Martin, and the user can edit or ignore it before sending manually.
Multi-device sync
Sync keeps context with you
Continue relationship context across Mac, Windows, iOS, and Web.
03 Partner
A practical message becomes shared context instead of disappearing in chat.
Alex and Sam use one shared space for household context, but private notes stay private on each person’s device.
Source
Sam sends a photo of a receipt and says: “This is the cleaner we liked. It did not trigger allergies.”
AI asks
Soulife proposes saving the useful part, not the whole conversation: product name, reason, and source photo.
Saved context
Saved to: Shared home space / Household preference. “Cleaner worked well and did not trigger allergies. Source: receipt photo from chat.”
Next moment
Two months later, AI can answer “what cleaner did we like?” with the source and the person who confirmed it.
09 Relationship Radar
Daily Radar picks only the 1-3 relationships that matter today.
A busy founder has many relationships, but does not want a task queue full of social reminders.
Source
Today’s signals: a friend deciding on a job, a parent with a medication check, and an old friend where repair needs space.
AI asks
Soulife asks: “Should today prioritize care, collaboration, or giving space?”
Saved context
Saved to: Relationship Radar. “Today prioritize one gentle medication check-in; hold the repair relationship for now.”
Next moment
Radar explains why one person appears and why two others are intentionally quiet.
10 Chat reminder
A selected chat fragment becomes a gentle reminder with source.
Roommates are handling apartment paperwork and want to remember details without turning chat into a project board.
Source
Chat: “Can you remind me Sunday night to send the signed form to the landlord? I might forget.”
AI asks
Soulife asks: “Should this be a reminder, a shared action, or a natural relationship touchpoint?”
Saved context
Saved to: Chat-extracted reminder. “Sunday night: remind them to send the signed form. Source message attached.”
Next moment
When the reminder appears, the source card can jump back to the original chat message.
Offline usable
Offline keeps the product dependable
Local context remains readable and actionable, then syncs later.
05 Reconnecting
A relationship goal changes the tone of the next suggestion.
Chris wants to repair an old friendship, but does not want AI to push aggressive follow-ups.
Source
Relationship goal: “Repair slowly. Do not force a long conversation unless they show interest.”
AI asks
Soulife uses the goal as a boundary when reviewing recent chats and avoids task-like reminders.
Saved context
Saved to: Jamie / Relationship direction. “Repair slowly; prefer short, no-pressure touchpoints.”
Next moment
Instead of “follow up today,” Soulife suggests: “Give it a few days. If you message, keep it light and specific.”
06 Relationship goal
A vague relationship gets a direction before it gets advice.
Elaine is deciding whether a former teammate should stay a light collaboration, become a closer friendship, or simply rest.
Source
Relationship note: “She was helpful before, but our recent chats are polite and shallow. I do not know where this relationship should go.”
AI asks
Soulife asks one narrowing question first: “Should this move toward maintaining, deepening, repairing, collaborating, or letting it rest?”
Saved context
Saved to: Relationship goal. “Light collaboration with boundaries; good for project or industry check-ins, not personal closeness.”
Next moment
The next AI draft suggests sharing one concrete industry note instead of pushing for a personal meetup.
07 Natural rhythm
Interaction frequency becomes a human rhythm, not a score.
Sofia and her older cousin talk a few times a month. The cousin recently mentioned visa paperwork stress.
Source
Signals: 3 chats in 2 months, 2 started by the cousin, and an open topic about paperwork due this week.
AI asks
Soulife asks: “Is this a light check-in this week, give it space, or make room for a deeper talk?”
Saved context
Saved to: Natural rhythm. “Lightly ask about visa paperwork once this week; no long call unless she opens up.”
Next moment
Relationship Radar shows a single suggestion with the reason: ask once, then stop unless she expands.
Verified loops
The landing page promises are backed by runnable product fixtures.
The same 10 scenarios above are executable inside Soulife’s hidden product fixture. It verifies source context, AI question, preview, confirm-save, target readback, reload persistence, and an Electron screenshot run.
01
Product promise registry keeps the marketing page and in-product fixture in the same order.
02
Web and Electron E2E tests click through every loop and confirm saved records survive reload.
03
The screenshot below is generated from the real Electron fixture, not a static illustration.
How AI fits
AI is a context operator, not a hidden reader.
Soulife is designed around client-side context selection and a preview-and-confirm loop. AI can suggest a memory, a relationship direction, or a next natural touch, but the user sees the source chain and confirms before anything becomes reusable context.
Read the privacy model 1 Source selected on device
A message, note, imported chat, or workspace entry gives the reason.
2 Workspace context
Soulife shows which private or shared space would receive the context.
3 Confirm before write
AI actions become real only after preview, confirmation, and a visible source link.
Start with one conversation, one space, or one relationship.
Use Soulife to keep private context useful without making your personal life feel like a task board.
Get the desktop preview