Fotoshi
 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
  | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Author: | Fotoshi Inc. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| License: | proprietary (software), CC BY 4.0 (user photos), and other open licences (some imports) (free of charge) | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Platforms: | Android, iOS, and Web | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Status: | Active | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Language: | multiple languages
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Website: | https://fotoshi.com | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Install: | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Give every place a photo | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Fotoshi is the photo-first platform to give every mappable place and object in the physical world a photo. Users capture one or many photos with the mobile app for Android or iOS, then select or add a place on the map to link the photos to, offering proof of existence and a depiction of the attributes. Photos are CC-BY with the goal of making them maximally reusable with credit to the author, and the long term vision is to enable every entity on the map to have detailed attributes, visual location improvement, and change logs over time grounded in photo evidence from the community worldwide.
Development history
Fotoshi was first announced by founder Edoardo Neerhut at FOSS4G Hiroshima and State of the Map Asia in Osaka in September 2026. During September and October 2026 beta testing for Android and iOS opened for the mobile app. The Android app launched on Google Play at the end of October 2026, followed by the iOS app on the App Store. Christopher Beddow joined the team in November 2026 as co-founder.
Support
Fotoshi has received NVIDIA Inception/AWS cloud storage and compute credits in support of hosting and processing photos.
Fotoshi Examples
- Review a restaurant's menu and appearance, and check its opening hours against current photos
- Share a panoramic coastal view
- Inspect steps; compare a ramp and path surface for accessibility
- Find a defibrillator and see where it is installed
- Find an ATM and see its surroundings
- Read an on-site information sign to learn about the place
- Use posted maps and directories to find your way; photograph floor plans too
- Potential community mapping project: photograph a business's storefront, name sign and entrance, then revisit it to document visible changes over time.
- Photograph other mapped objects too, such as benches, hydrants, EV chargers and trail markers.
Computer Vision
Fotoshi has plans to work on two pillars with a computer vision focus:
- on-device algorithms to assist users in choosing which nearby object the photo depicts, or which category and name represent a newly added place
- server side place intelligence to find new attributes of a place or object, such as its category, subcategory, material, accessibility, updated name, services or products offered, brand, purpose, condition, contact information, and more.
The team is seeking partners in academia and research institutions interested to test novel uses of Fotoshi CC-BY images as a spatial, temporal, and visual dataset, and welcomes contact from interested people and organizations.
Use in OpenStreetMap
Mappers can currently use the prototype integration to OpenStreetMap iD editor at https://osm.fotoshi.com. This shows the Fotoshi photo overlay, which enables seeing places that have a gallery of photos, marked as places already on OSM and those that are not on OSM (either user created or sourced from Overture, and now photo-verified).
Users can reference the photos to add new attributes to a place such as reading its opening hours or accessibility status, and click a button to add the Fotoshi photo ID as image:fotoshi=$id in the object's tags. It may be possible in the future to use a specific Fotoshi key that links to the Fotoshi place ID which can be thought of as a gallery ID for that associated OSM object ID, enabling linking of dozens or more photos.
As Fotoshi develops its capabilities, computer vision extracted information such as contact information, street address, types of cuisine, accessibility, and other attributes will be available to add to the OSM object when presented via the Fotoshi overlay layer and photo viewer. Photos may also give evidence of a different spelling of a name, a better category than what OSM carries, or evidence overriding existing attributes, which the user can update with their own judgment. The community can help suggest future improvements for iD editor and ideas on how to best integrate to JOSM and other editing tools as well.
Feature requests
Members of the OSM community have made some feature requests:
- Allow importing photos from the gallery, while balancing risk of anyone uploading photos without GPS, clear ownership and license provenance, and high volume of low quality or outdated photos
- Allow OSM users to enable a direct changeset to OSM when adding new objects to the map via authenticated linking to their OSM account and permission to submit changesets on their behalf
The Fotoshi team is open to consider features like this while balancing quality contributions into Fotoshi, and from Fotoshi into other platforms. The developer documentation may offer what independent developers need to make their own implementation of some feature requests as well.
Developer API
The developer page describes the STAC API as providing “the same public photo observations in STAC format for GIS and STAC clients.”[1] The GeoJSON Photos & Features API supports contributor-based photo searches, finding objects with photo coverage, and looking up Fotoshi FIDs, OSM identities, and Overture GERS IDs. The page also offers MCP access for AI clients through Fotoshi sign-in. Small STAC searches are anonymous; API keys provide larger account allowances.
OSM community outreach
The Fotoshi team aims to collaborate with fellow OpenStreetMap members to provide training, organizing workshops, and coordinate mapping projects, where Fotoshi can be a useful field data tool or source of open data. The Fotoshi team also aims to attend as many OSM events as possible.
Privacy
- Stored by Fotoshi: Original photos, including embedded EXIF, plus extracted camera, capture and pose metadata. Owners can download or export their originals.[2]
- Served publicly: Place/object IDs, category, attribution, licence, location and dates, camera direction and dimensions.
- Not served publicly: Raw EXIF, including camera-owner names or serial numbers; raw sensor or GNSS records; uploader-supplied original filenames.
Fotoshi evaluated YOLOv11 but chose not to use it. Its face-redaction pipeline uses OpenCV’s YuNet detector, with additional models providing verification and person-context checks. Public image renditions apply a pixelation-and-blur treatment to regions selected for redaction. Special handling is given to photos linked to statues, artwork, sculptures and other categories at high risk of a false, non-human face blur.
Fotoshi moderates uploads and user reports. Users can report obscenity, violence, privacy violations, faces that failed to be blurred, wrongly blurred objects, and request license plate blurs.
License
“Photos uploaded by Fotoshi contributors via the Fotoshi apps are published under the Creative Commons Attribution 4.0 International licence (CC BY 4.0).”[3]
Photos imported from open photo collections may use a different license, equivalent in openness to CC-BY or more open. Check the licence shown on each photo.
- Jeonju photo from the Korea Tourism Organization import — KOGL Type 1; API attribution fields: photographer 김찬영; source organization: Korea Tourism Organization.
- When reusing a photo, preserve contributor and source attribution, and identify Fotoshi as the source where applicable.
- Public photos may be used to derive facts and map data for OpenStreetMap under the OSM Foundation contributor terms and the ODbL.
See also
- Photo mapping - Using pictures to describe features in OSM
- Photo linking - Create links between OSM features and external pictures
- Geotagging Source Photos - Add coordinates information on non-geolocated pictures
- Street-level imagery services - Alternatives
- Wikimedia Commons - Photos can be exported from the Fotoshi photo viewer using its export button.
- Fotoshi developer documentation
- Fotoshi for Android (Google Play) and Fotoshi for iOS (App Store)
Map a Post Office with Fotoshi - quick introduction to the user experience