🍁MapleLeaf Community🌐 Interactive Platform Guide
🧭 Welcome to MapleLeaf Community · Interactive Tutorial & Reference

Turn Treasured Photos, Stories & Records into a Living AI Archive

Browse at your own pace. Start with the 4 big walkthroughs below, test-drive features safely in the interactive sandbox, or search our complete encyclopedia of archive tools.

πŸ‘οΈ Choose Your Guide Lens: (Click any role to reorder the walkthroughs below)
Complete Platform Tour: Explore everything MapleLeaf Community can do β€” from conversational search and oral history recording to knowledge graphs, face scanning, and custom entity schemas.
ZONE 1 Β· STEP-BY-STEP WALKTHROUGHS

What would you like to learn first?

Click any large card to open its 4-step visual guide
πŸ’¬
⏱️ 2 min walkthrough

Learn how to chat with your archive

Ask everyday questions, explore original photos & audio, and discover companion cards

β–² Hide Steps

You never need complex search filters. Just ask a question like you would ask a family historian, and the AI weaves together stories cited directly to original photographs, letters, and recordings.

πŸ’¬
Step 1 of 4

Ask natural questions or click Starter Chips

Type anything in plain English β€” a person’s name, a town, a decade, or "Tell me a story about Grandma’s childhood." Not sure where to start? Click any of the curated Starter Query chips on the welcome screen.

✨ Key Highlights: Try: "Show me photos from the 1960s" or "Who lived at the old homestead?"
πŸ’‘ Tip: You can ask follow-up questions in the same conversation β€” the archive remembers your context.
ZONE 2 Β· INTERACTIVE PRACTICE SANDBOX

Try It Yourself Before Touching Live Archive Files

πŸ§ͺ 100% Safe Practice Sandbox Β· Nothing You Click Here Modifies Live Data

Hands-On Practice Simulator

Click around to experience how tagging a photo, recording an oral history, or inventing a custom entity immediately enriches the archive.

Step 1: Curate Sample PhotoSimulates AnnotationPanel.js
πŸ“Έ
Summer_Picnic_Hawkesbury_1968.jpg
Family picnic table by the wooden boat jetty, Sunday afternoon
#1960s#summer-picnic#hawkesbury
Step 2: Live Effect in Archive ChatUpdates immediately as you toggle!
πŸ€– ARCHIVE AI RESPONSE PREVIEW

In Summer 1968, the family gathered at Hawkesbury River Cottage for the 1968 Summer Reunion. Pictured in Summer_Picnic_Hawkesbury_1968.jpg: Margaret "Peggy" Godbolt.

πŸƒ Companion Cards Triggered Beside This Answer (2):
πŸ‘€
Margaret "Peggy" Godbolt1934 – 2018
Aliases: Grandma Peggy, Peggy Anne Β· Matriarch Β· Mother of 4
Born in Parramatta, known for her Sunday roast gatherings, community choir leadership, and meticulous photo albums.
🎲 Random Source
πŸ—ΊοΈ
πŸ“ Hawkesbury River Cottage, NSW
Coordinates: -33.5420, 151.1980 Β· Interactive Google Map Embedded
Satellite / Street View β†—
ZONE 3 Β· COMPLETE FEATURE REFERENCE & LOOKUP

Look Up Every Aspect of Using & Managing the Archive

πŸ”
πŸ‘€

People, Aliases, Relationships & Family Tree (🌳)

How to manage people records, maiden names, nicknames/aliases, automatic reciprocal family relationships, duplicate merging, and the interactive Family Tree.

⚑ At a Glance
  • Aliases teach the AI that "Bob", "Robert", and "Grandpa Bob" are the exact same person.
  • Adding "Alice is Parent of Ben" automatically adds "Ben is Child of Alice".
  • Merging two duplicate people combines their bios, aliases, relationships, and tagged photos without losing anything.
β–Έ Structured Person Fields & Why Aliases MatterClick to toggle
Each Person record in Admin β†’ 🧠 Knowledge β†’ πŸ‘€ People stores:
β€’ Full Display Name, Given Name, Middle Name(s), Family Name (Surname), and Name at Birth / Maiden Name.
β€’ Aliases / Nicknames: Comma-separated variations (e.g., *Grandma Peggy, Margaret Anne, M.A. Godbolt*). Whenever the AI rebuilds its knowledge index, every alias is mapped to that person so searches for any nickname find all of their documents and photos.
β€’ Birth Year & Death Year, Gender, Biography / Notes, and Portrait Photo.
β–Έ Automatic Reciprocal RelationshipsClick to toggle
When editing a Person, you can connect them to other people in the archive using relationship types:
β€’ Parent / Child (automatically creates the inverse child or parent link on the other person)
β€’ Spouse / Partner (automatically creates the reciprocal spouse link)
β€’ Sibling (automatically creates the reciprocal sibling link)
β€’ Other relative or associate
These relationships power both the Person Companion Cards in Chat and the interactive 🌳 Family Tree visualization.
β–Έ Merging Duplicate People Safely (πŸ”€)Click to toggle
When AI entity extraction scans letters and transcripts, it might create *"Uncle Arthur"* and *"Arthur Godbolt"* as two separate entries.
1. In 🧠 Knowledge β†’ πŸ‘€ People, click πŸ”€ Merge Duplicate on the record you want to keep.
2. Select the duplicate person record to merge into it.
3. The archive transfers all file links, combines aliases and relationships, and removes the duplicate cleanly.
β–Έ Interactive Family Tree (🌳)Click to toggle
Available in Admin β†’ 🌳 Family Tree (and linked from Person cards):
β€’ Automatically lays out generations from oldest ancestors at the top to youngest descendants at the bottom.
β€’ Supports pan, zoom, and click-to-inspect: clicking any person node opens their portrait, lifespan, relationships, and a list of archive files they appear in.
πŸ“

Places, Coordinates & Interactive Google Maps

How Places work, adding Latitude/Longitude coordinates, triggering embedded satellite/street maps in Chat, and saving locations with one click.

⚑ At a Glance
  • Any Place with Latitude & Longitude coordinates automatically renders an interactive Google Map card in Chat.
  • Archivists can click "πŸ“ Save Location to Archive" directly inside Chat when the AI identifies a place.
  • Places can represent towns, homesteads, churches, schools, battlefields, or countries.
β–Έ Place Hierarchy & Structured FieldsClick to toggle
In Admin β†’ 🧠 Knowledge β†’ πŸ“ Places, each place includes:
β€’ Place Name (e.g., *St. John’s Church, Parramatta* or *The Old River Homestead*)
β€’ Locality / Town, Region / State / Province, and Country
β€’ Latitude & Longitude (decimal coordinates, e.g., -33.8151, 151.0011)
β€’ Description / Historical Context and Files Linked count
β–Έ Interactive Map Companion Cards in ChatClick to toggle
When a chat answer references a Place that has coordinates (or when the AI resolves a geographical location from the archive), an interactive Map Companion Card appears beside the chat response:
β€’ Switch between Roadmap and Satellite imagery.
β€’ Click Open in Google Maps β†— to explore street view or directions.
β€’ If the AI inferred coordinates that aren’t saved in your Knowledge Directory yet, Archivists see a one-click πŸ“ Save Location to Archive button right on the map card!
πŸŽ‰

Events & Automatic Context Linking

How Events connect People, Places, Dates, and Source Files together β€” and how tagging an Event in Chat automatically links its participants and location.

⚑ At a Glance
  • An Event connects a Date, a Place, and multiple People (Participants) into one historical moment.
  • When you tag a photo with an Event in the Annotation Panel, the archive can automatically attach the Event’s Place and People!
β–Έ What Makes an Event PowerfulClick to toggle
In Admin β†’ 🧠 Knowledge β†’ πŸŽ‰ Events, an Event represents a specific occasion β€” such as a *1958 Golden Wedding Anniversary*, *The 1974 Flood*, *Annual Summer Camping Trip*, or *Moving to Australia*.
Each Event stores:
β€’ Event Name & Date / Date Range
β€’ Associated Place (linked directly to a Place record)
β€’ Involved People / Participants (linked to Person records)
β€’ Description & Linked Archive Files
β–Έ Smart Auto-Propagation When AnnotatingClick to toggle
When you open ✏️ Annotate on a photo or document and select an Event, the annotation system knows which People and Place belong to that Event. This saves huge amounts of time when curating photo albums from weddings, reunions, holidays, or community milestones.
✨

Custom Entity Types & Custom Fields (Schema Builder)

Everything about creating custom entity categories, using AI to suggest schemas, all 8 custom field types, card subtitles, custom icons/colors/shapes, and hiding built-in types.

⚑ At a Glance
  • Create unlimited custom entity types (e.g., Family Recipes, Heirlooms, Boats, Pets, Military Units, Awards).
  • Click "✨ Suggest Schema with AI" and type a category name β€” the AI designs the fields, icon, and description for you.
  • Choose from 8 field types: Text, Long Text, Number, Date, Select (Dropdown), Tags, URL, and Entity Reference.
  • Mark up to 3 fields as "Card Subtitle" so they appear prominently on Companion Cards in Chat.
β–Έ Creating a Custom Entity Type (Step-by-Step)Click to toggle
1. Go to Admin β†’ 🧠 Knowledge and click the + Custom Entity Type button in the sub-tab bar.
2. Enter a Singular Label (e.g., *Family Recipe*) and Plural Label (e.g., *Family Recipes*).
3. *(Optional)* Click ✨ Suggest Schema with AI β€” the AI will automatically propose tailored fields (like *Ingredients*, *Prep Time*, *Originating Relative*, *Occasion*), pick an emoji icon, and write a description!
4. Customize the Emoji Icon, Badge Color, and Entity Graph Node Shape (● Circle, β–­ Rounded Rect, β—† Diamond, ⬑ Hexagon, β–² Triangle, or β˜… Star).
β–Έ All 8 Custom Field Types ExplainedClick to toggle
Every custom entity always includes a Name, Aliases, Description, and Photo, plus as many Custom Fields as you want:
β€’ `text` (Short Text): Single-line text (e.g., *Model Number*, *Regiment*, *Callsign*).
β€’ `textarea` (Long Text): Multi-paragraph text (e.g., *Cooking Instructions*, *Provenance Story*).
β€’ `number` (Number): Numeric values (e.g., *Year Built*, *Tonnage*, * Valuation*).
β€’ `date` (Date): Historical or calendar dates (e.g., *Date Awarded*, *Voyage Date*).
β€’ `select` (Dropdown): Predefined choices you list as comma-separated options (e.g., *Condition: Mint, Good, Fair, Fragile*).
β€’ `tags` (Tag List): Multi-value chips (e.g., *Materials: Silver, Oak, Brass*).
β€’ `url` (External Link): Clickable web link (e.g., *Museum Registry URL*, *External Reference*).
β€’ `entity_ref` (Entity Reference): Links directly to another entity in your archive (e.g., *Created By β†’ Person*, or *Kept At β†’ Place*)!
β–Έ Card Subtitles & Hiding Built-In TypesClick to toggle
β€’ Card Subtitle Checkbox: Toggle Card Subtitle on up to 3 fields. When a Custom Companion Card appears in Chat, those 3 fields are highlighted right underneath the entity’s title.
β€’ Hiding Built-In Types: Don’t need *People*, *Places*, or *Events* in a specialized archive? You can hide any built-in entity type from the Knowledge bar and restore it anytime with one click.
🧠

Knowledge Graph, Co-occurrence & AI Extraction

How the archive’s hybrid brain combines vector RAG with structured Markdown knowledge files, AI Entity Suggestions (πŸ€–), Co-occurrence links, and the Entity Graph (πŸ”—).

⚑ At a Glance
  • The archive uses a Hybrid Brain: raw files (photos, PDFs, transcripts) + auto-generated structured Knowledge Markdown files.
  • "πŸ€– Suggest Entities" scans your indexed documents and proposes new People, Places, and Events for your approval.
  • "πŸ”„ Rebuild Knowledge Base" compiles your latest entity edits and re-indexes them into the Vertex AI RAG engine.
β–Έ How the Hybrid Knowledge Brain WorksClick to toggle
When you add or edit People, Places, Events, or Custom Entities, the platform compiles structured knowledge documents (knowledge_people.md, knowledge_places.md, knowledge_events.md, and custom type manifests) and indexes them directly alongside your original files in the Vertex AI RAG corpus.
That means when a user asks *"How is Margaret related to Arthur, and where did they live?"*, the AI reads both your structured family knowledge graph AND the original letters/transcripts!
β–Έ Using "πŸ€– Suggest Entities" (Automated AI Discovery)Click to toggle
Instead of typing every person and place from scratch:
1. Open Admin β†’ 🧠 Knowledge and click πŸ€– Suggest Entities.
2. The AI reads through your archive’s transcripts and documents to extract candidate People, Places, Events, and Custom Entities (complete with confidence scores and source snippets).
3. Review the suggestions: click βœ“ Approve on the ones you want to keep, edit any details, or dismiss false positives.
β–Έ Exploring the Interactive Entity Graph (πŸ”—)Click to toggle
In Admin β†’ πŸ”— Entity Graph:
β€’ Every entity is rendered as a color-coded, shape-coded node (sized by how many archive files reference it).
β€’ Lines (edges) connect entities that share relationships or appear together in the same photos, documents, or recordings (Co-occurrence).
β€’ Filter by entity type, search for a specific node to highlight its immediate neighborhood, or drag nodes to explore clusters in your collection.
πŸ“·

Visual Face & Entity Scanner (PersonScannerModal)

How to crop a person’s face from one photo and let Gemini Vision scan your untagged photo collection to find matching appearances.

⚑ At a Glance
  • Works directly from any Person card in Admin β†’ 🧠 Knowledge β†’ πŸ‘€ People (click "πŸ“· Scan").
  • Includes a built-in interactive Face Cropper so you can isolate one face from a group photo.
  • You stay in 100% control: review candidate matches side-by-side and check only the photos you want to link.
β–Έ Step-by-Step: Scanning the Archive for a PersonClick to toggle
1. Go to Admin β†’ 🧠 Knowledge β†’ πŸ‘€ People and find the person you want to scan for.
2. Click the πŸ“· Scan button on their row.
3. Choose or Crop a Reference Portrait: If they don’t have a close-up portrait yet, pick any group photo from the archive or upload an image, then drag the crop box over their face using the interactive Face Cropper. That cropped face can also be saved as their official profile portrait!
4. Configure Scan Scope: Choose whether to scan *Untagged images only* (fastest) or *All archive images*, and pick your batch size.
5. Review & Confirm: Gemini Vision compares the reference face against your archive photos, returning High and Medium confidence matches with visual reasoning notes. Check the confirmed photos and click Link Selected Files.
✏️

In-Chat Annotation, Batch Annotate (πŸ“) & Taxonomy Tags (🏷️)

How to annotate single files in Chat, tag 50 files at once in Batch Annotate, record spoken annotations on photos, and manage color-coded taxonomy tags.

⚑ At a Glance
  • Use In-Chat "✏️ Annotate" when exploring conversationally; use "πŸ“ Batch Annotate" when curating whole folders of photos.
  • In Batch Annotate, toggle "Untagged only" to rapidly work through newly indexed files up to 50 at a time.
  • In "🏷️ Tags", you can customize tag colors, descriptions, and merge duplicate tags across the whole archive.
β–Έ In-Chat Annotation Panel vs. Batch Annotate (πŸ“)Click to toggle
β€’ In-Chat Annotation (`✏️ Annotate`): Best for deep, contextual curation on a single photo or document β€” lets you add People, Places, Events, Custom Entities, Dates, detailed Notes, AND record spoken audio/video commentary (πŸŽ™οΈ Record tab) while viewing the file.
β€’ Admin Batch Annotate (`Admin β†’ πŸ“ Batch Annotate`): Best for high-speed bulk organization. Filter by file type (Images, Audio, Documents), search by filename, or toggle Untagged only. Select up to 50 files at once and apply shared People, Places, Events, Tags, Date, and Notes in a single click!
β–Έ Managing Your Tag Taxonomy (🏷️ Tags)Click to toggle
In Admin β†’ 🏷️ Tags:
β€’ View every topic tag used across the archive alongside its usage count.
β€’ Assign custom badge colors and descriptions so explorers understand what each tag means.
β€’ Merge Tags: Consolidate synonymous tags (e.g., merge #ww2 into #world-war-ii) β€” every file carrying the old tag is updated automatically.
πŸŽ™οΈ

Storyteller Portal, Catalyst Prompts (πŸ’­), Review Queue & User Roles (πŸ‘₯)

Managing the Storyteller experience, customizing Catalyst Prompts, moderating contributions, editing transcripts, and inviting users as Viewer, Archivist, ArchiveSource, or Both.

⚑ At a Glance
  • Users can hold Archivist rights, ArchiveSource (Storyteller) rights, or BOTH simultaneously.
  • Archivists can customize the Catalyst Prompt Bank (πŸ’­) with family-specific or community-specific questions.
  • Spoken recordings are transcribed automatically by AI, and Archivists can polish the transcript text in "πŸŽ™οΈ Recordings".
β–Έ Understanding User Roles & Invite Links (πŸ‘₯ Users)Click to toggle
In Admin β†’ πŸ‘₯ Users, you can invite relatives or community members by email or generate a shareable invite link with exact role checkboxes:
β€’ Viewer (Explorer): Can sign in, chat with the archive, view photos/maps/trees, and submit questions via *❓ Ask a Question*.
β€’ Archive Source (Storyteller): Defaults to the simple, confidence-first Contribution Portal (/tenant/{id}?source=1) to record stories, answer questions, and upload files, with a one-click button to switch to Chat.
β€’ Archivist (Curator): Can annotate files in Chat, save map locations, and open the full Admin Dashboard (/admin/{id}) to manage the Knowledge Graph, Custom Schemas, and Batch Annotations.
β€’ Both (`Archivist` + `ArchiveSource`): Grants full access to both the Storyteller Portal and the Archivist Curation tools!
β–Έ Customizing Catalyst Prompts (πŸ’­) & Community Questions (❓)Click to toggle
β€’ πŸ’­ Catalyst Prompts: Seeded with 60+ thoughtful oral-history questions across 6 categories (*Childhood*, *Family Traditions*, *Parents & Grandparents*, *Career*, *Historical Events*, *Life Lessons*). Archivists can add custom questions specific to your family or community!
β€’ ❓ Questions: When a viewer clicks *❓ Ask a Question* in Chat, it lands here and in the Storyteller Portal so elders or archivists can record or write an authoritative answer.
β–Έ Moderating Contributions (πŸ“₯ Review) & Polishing Transcripts (πŸŽ™οΈ Recordings)Click to toggle
β€’ πŸ“₯ Review Queue: Inspect newly submitted stories, recordings, and file uploads before or after they enter the corpus.
β€’ πŸŽ™οΈ Recordings: Listen to every oral history recording, read the AI-generated transcript, and edit any spelling of family surnames or town names so the RAG search index is 100% accurate.
🌊

Archive Billing Streams, $0.25 Flag-Fall & Family vs. Open Routing

How real-time LlamaPay billing streams work, the $0.25 USD eq. ($0.38 AUD) stream initiation flag-fall, Archive SAFE + LlamaPay Pool balances, and why Closed/Family Archives route 100% directly to the Platform Treasury SAFE.

⚑ At a Glance
  • Every archive includes 20 free daily queries before a real-time LlamaPay stream starts.
  • Starting an engagement stream includes a $0.25 USD (or currency equivalent, e.g. $0.38 AUD) upfront flag-fall batched into the stream creation transaction to cover Arbitrum L2 start/stop gas.
  • Closed / Family Archives route 100% of all payments (streams, flag-falls, storage, face scans) directly to the Platform Treasury SAFE (0xB449...63af).
  • Open Community Archives route 95% of ongoing engagement streams to the Archive Owner and 5% to the Platform Treasury SAFE.
β–Έ Two-Part Transparent Metering (Model 1 Engagement + Model 2 Storage)Click to toggle
MapleLeaf uses transparent, pay-for-what-you-use on-chain streaming on Arbitrum One:
β€’ Model 1 β€” Active Engagement Stream (after 20 free daily queries): Default $2.00 AUD / $1.50 USD per 30 minutes (~$0.0011 AUD/sec) while actively exploring the archive, plus a $0.25 USD (or currency equivalent: $0.38 AUD, Β£0.20 GBP, €0.23 EUR, $0.34 CAD, $0.41 NZD, $0.34 SGD, 2.60 kr SEK) upfront stream flag-fall.
β€’ Model 2 β€” Unified Archive Storage Stream: Default $10.00 AUD / $6.50 USD per GB / month, streamed continuously per second from the Archive Owner’s SAFE + LlamaPay Pool to cover Google Cloud Storage and Vertex AI RAG vector index embeddings.
β–Έ Why There Is a $0.25 USD Equivalent Stream Initiation Flag-FallClick to toggle
Starting and stopping an on-chain LlamaPay stream on Arbitrum One requires two Gnosis Safe execTransaction calls (~246k gas to start + ~136k gas to stop) plus initial AI context warming.
β€’ Without a flag-fall, a very short session (e.g. 15–30 seconds) would cost more in blockchain transaction overhead than the streamed value transferred.
β€’ The $0.25 USD eq. ($0.38 AUD) flag-fall is batched atomically inside the MultiSendCallOnly stream creation transaction (0 extra blockchain transactions) and routed directly to the Platform Treasury SAFE (`0xB449...63af`).
β–Έ Closed / Family Archives vs. Open Community ArchivesClick to toggle
β€’ Closed / Family Archives (`community_private` / invite-only): Closed family archives are not commercial creator channels. 100% of all payments β€” including the $0.25 USD eq. flag-fall, Model 1 engagement streams, Model 2 storage streams, and AI photo face scan charges β€” go 100% directly to the Platform Treasury SAFE (`0xB449950D5659a809881059ccE8fE1995b25c63af`) to pay for storage, vector indexing, and AI compute.
β€’ Open Community Archives (`community_open`): Public creators and historical societies receive 95% net of ongoing Model 1 engagement streams, with 5% platform fee + the $0.25 USD eq. flag-fall routed to the Platform Treasury SAFE.
🍁 MapleLeaf Community

Sign in to MapleLeaf

Enter your email for an instant passwordless sign-in link, or continue with Google.