Skip to content
finwave Docs

Onboarding Overview

In this guide you will learn:

  • The four stages of onboarding: Discovery, Manifesting, Review, and Sync
  • What each stage does and what you need to do at each step
  • How local and cloud-stored data are handled differently

What onboarding does

Section titled “What onboarding does”

Onboarding is the process of getting your existing image collection into finwave. The desktop app guides you through a structured pipeline that organizes your raw files into encounters, verifies the data, and uploads everything to the server where the ML pipeline processes it.

The pipeline

Section titled “The pipeline”

1. Discovery

Section titled “1. Discovery”

Scan your image directories to build a file inventory. Discovery extracts metadata from every image (EXIF dates, GPS, camera info, IPTC creator) and analyzes folder structures and spreadsheets for additional encounter data.

  • Local directories — Select folders on your machine or external drives. Scanning runs locally.
  • Azure blob storage — Connect a storage account and scan blob containers remotely. A cloud worker processes the files without downloading them to your machine.

Learn more about Discovery

2. Manifesting

Section titled “2. Manifesting”

Configure how your files are grouped into encounters. A manifest defines the rules for extracting date, location, photographer, and individual IDs from your file metadata.

  • Choose a grouping strategy (folder, filename, time proximity, or fusion)
  • Configure field source mappings
  • Resolve photographer names to finwave user accounts
  • Preview the results and refine until the grouping looks correct

Learn more about Manifesting

3. Pre-Sync Review

Section titled “3. Pre-Sync Review”

After approving a manifest, it is materialized into concrete encounter records. The pre-sync review lets you inspect each encounter:

  • Verify data completeness (date, location, GPS, photographer)
  • Assign licenses (Public, Public+Attribution, or Private)
  • Assign organizations
  • Approve or deny individual encounters
  • Edit encounter data inline if corrections are needed

Learn more about Pre-Sync Review

4. Sync

Section titled “4. Sync”

Upload approved encounters and their images to finwave. The sync engine:

  • Creates encounter records on the server (or finds existing ones via dedup)
  • Uploads images via SAS URIs (local files) or blob-to-blob copy (Azure files)
  • Tracks progress per encounter and per image
  • Handles retries on transient errors automatically
  • Triggers the ML pipeline for each uploaded image

Learn more about Sync

Resumability

Section titled “Resumability”

Each stage saves its state. You can close the app and resume later:

  • Scans persist in the local database
  • Manifests are saved as you edit them
  • Materialized encounters persist across sessions
  • Sync progress is tracked per image — interrupted syncs resume where they left off

Local vs. cloud data

Section titled “Local vs. cloud data”
Local directoriesAzure blob storage
DiscoveryScans locally, fastCloud worker, may take hours for large datasets
MetadataFull EXIF/IPTC extractionHeader-only download (128KB per image)
Sync uploadDesktop uploads via SAS URIServer-side blob-to-blob copy (fast, no local bandwidth)
Section titled “Related”