User Guide

Each saved item is an entry. An entry has a title, body, time, privacy setting, optional categories, and optional media.

The operating loop is Capture, AI review, human review, save, organize, route, and reuse. The original Conscious Log entry remains available after routing.

The app is multiuser and multitenant. Workspace membership, roles, teams, sharing choices, and tenant boundaries control access.

A pending invitation becomes accepted after the first matching sign-in. Admin can see Pending, Accepted, Active, or Expired status and can resend a pending invitation.

The left navigation opens Today, Timeline, Patterns, LifeTilo routes, User Guide, and Admin. The top bar provides private-log search, notifications, Import, and New entry.

Private by default means a new entry starts as Only me. The author must deliberately select people, teams, or the organization before other members can read it.

Text: AI corrects spelling, grammar, clarity, and coherence, shows the corrected draft, then saves the approved text.

List: each line becomes a checklist item and keeps its checked state.

Image: mobile users can take a picture or choose one. AI reads visible text, understands the picture, proposes a title, and places a description in the caption area for review.

Drawing: the drawing canvas is stored as image media with the note text.

Audio: the browser records audio and may create a live transcript when speech recognition is supported. The transcript becomes the entry body.

Check-in: stores the selected energy score, current feeling, and note as a restricted check-in entry.

Text cleanup uses the configured editor instruction and returns corrected text only. The user reviews the result and presses Save corrected text.

Image analysis uses a vision-capable OpenAI model. It extracts visible wording, describes what the image communicates, and generates a short title.

PDF, DOCX, DOC, TXT, and Markdown attachments can be summarized. The summary becomes the body and the original file remains attached.

Admin selects the default model and purpose-specific models under Model routing. AI usage shows calls, tokens, estimated USD cost, date range, and per-user totals.

Each row shows the generated or edited title, body preview, time, privacy, categories, route state, and media status.

Edit entry changes the title and body. If media is attached, the editor also offers Remove picture or Remove attachment.

Categories can be added or edited without leaving the card. Archive hides the record from active lists while keeping it recoverable in Timeline.

Try AI analysis reruns media review when setup or processing previously failed.

Search matches title, body, attachment filename, and category. From and To dates restrict the result set, and Sort changes chronological order.

Filters show All entries, Notes, Check-ins, Routed, or Archived. Pinned cards remain at the top.

Long card text is shortened. Select the blue See More link to open a half-screen reader with scrolling, then choose Edit entry when changes are needed.

Card controls include Pin, Categories, Background, Reminder, Collaborator, Archive, and More. More contains Edit, Make a copy, Archive or Unarchive, and Delete.

An entry can have multiple categories such as Project, Idea, Brainstorming, or App DEV. Categories are tenant-scoped and can be created or renamed from the picker.

Pin keeps a card above unpinned results. Background changes the visual card treatment. Reminder stores a date, time, and optional location.

Make a copy creates a separate private entry. The copy can be edited and routed independently.

Only me limits access to the entry owner. People or teams requires at least one active workspace member or team. Organization makes the entry visible within the tenant.

Teams are created in Admin by selecting a team name and active member names. Sharing with a team uses that managed membership list.

Workspace roles currently include Owner, Workspace Admin, Collaborator Editor, Collaborator Viewer, and Auditor-style access where configured. Owners and permitted editors can manage records within their scope.

Tenant boundaries prevent one organization from automatically accessing another organization’s records.

Task receives the Conscious Log title as the task title, the body as notes, and related media as an attachment when supported.

Memory receives the title and body as reusable personal or organizational memory. Learning Card receives the title and explanatory body. Content receives the title and body for later publishing work.

The source card shows its destination and a route count. Repeated routing is recorded, and Timeline can filter records that have or have not been routed.

The routing confirmation is the human approval point. Review the title and body before approving the destination.

Supported import formats shown in the app include CSV, JSON, TXT, and Markdown. The mass-import workflow is designed around Title, Entry or Body, and Category columns.

CSV should use a header row such as title, entry, category. Multiple categories can be separated according to the preview guidance. JSON should contain an array of entry objects.

The import page explains supported structure and file limits, parses the file into a preview, reports rejected rows, and requires confirmation before writing entries.

Google Sheets should be exported as CSV before upload. PDF and Word files are supported as single entry attachments for AI summarization, not as the structured bulk-import format currently shown.

413 while processing a file means the request exceeded the accepted server size. The current uploader uses chunked transfer for larger files; retry after refreshing and confirm the file is within the displayed limit.

AI processing needs setup means the OpenAI API key or model route is missing. Admin must configure the environment and Model routing.

OpenAI API credits are exhausted means billing credit is unavailable in the OpenAI Platform account connected to this tenant. Add credits, then select Try AI analysis.

AI key needs attention indicates an invalid or revoked key. AI connection failed indicates a network or provider response problem. Rate limited means wait briefly and retry.

If an entry saves without appearing, clear Timeline filters, check Archived, and search the generated title or body wording.

A pattern names the observation, the number of source entries, and a confidence percentage.

Users can review the supporting entries and should correct or dismiss an observation that does not match their situation.

Patterns are guidance. They do not change, route, share, archive, or delete an entry without a separate user action.

LifeTilo receives approved Task, Memory, Learning Card, or Content records from Conscious Log.

The route preserves the source capture ID and destination metadata for audit and repeat-route tracking.

Google-based sign-in is an identity service. It is separate from importing Google Sheets content, which currently requires CSV export.

Future API documentation must define authentication, scopes, pagination, rate limits, idempotency, webhook signatures, retries, and redacted logging before publication.

Release-note entries should include version and date, changed behavior, affected roles, privacy or security impact, required action, and updated guide links.

Short tutorials should focus on one outcome, show the starting screen, list exact clicks, include captions or a transcript, and link back to the written procedure.

Guide owners should review search terms and support failures monthly, then replace outdated screenshots after interface changes.

Current guide evidence uses safe demo screens captured on August 22, 2026.

People: invite members, assign roles, see invitation status, resend access, and verify acceptance after matching sign-in.

Teams: create named groups and select active members. Teams become available in entry sharing.

Capabilities: review tenant feature availability and RBAC boundaries. Model routing: select default and purpose-specific OpenAI models and reasoning effort.

AI usage and AI call log: review call counts, token estimates, date-range USD cost, per-user usage, latency, model, purpose, and failures without copying private entry text into the admin report.

Import: review supported formats and commit tenant-safe bulk records. Audit log: review actor, action, resource, outcome, time, and redacted metadata.

Governance: use least privilege, review pending invitations, audit role changes, inspect AI failures and spend, and test recovery procedures on a recurring schedule.