Food Ecosystem

Store shopping

The ask behind this page: point a phone camera at a product on a store shelf and get back something useful, rather than trusting the packaging alone. 9 camera and barcode product scanners researched here, 2 of them open source with a verified repository. The rest are proprietary, several with real gaps for an Austrian shopper: US-only coverage, an unconfirmed privacy story, or a database that barely touches European retail assortments.

What exists

OSS open source, license and repository verified. proprietary closed source, license field left null. Sources are tiered A (repository, license text, official data-protection notice), B (independent journalism or NGO analysis), C (the vendor's own page or an app-store listing).

ServiceOSS / licenseData sourcePrivacy modelAustria / Vienna fit
Yuka food, cosmeticsproprietaryOwn proprietary product database, seeded from Open Food Facts and Open Beauty Facts in 2017 to 2018, then layered with a proprietary contribution-verification system since January 2018; around 700,000 to 800,000 products as of December 2019.Free to scan with no account required for first use, but server-side handling of scan activity is not independently verified here: Yuka's own privacy policy documents collecting account details (name, email, password), device data (device ID, model, manufacturer, OS and version), IP address and, only with explicit permission, location; it does not state whether product analysis runs on-device or server-side, and it does not support the narrower claim that only anonymised lookups are sent to Yuka's servers.Confirmed live in Germany and Switzerland (DACH neighbours) since 2021, but Austria was not on its own country launch list as of early 2020; since it draws on the same EU-wide barcode data as Open Food Facts, scanning products bought in Austria likely still resolves, just without an official Austrian rollout.
Open Food Facts foodOSS AGPL-3.0 (server and mobile app code); Open Database License / Database Contents License (the product data itself)Own collaborative, crowdsourced database (Product Opener server plus the Android, iOS and web apps), with nightly full-database dumps and exports published for offline use.Scanning and browsing require no account; contributing edits or photos does. The server, app and database dumps are all self-hostable, so a fully offline, no-telemetry deployment is a real option, not just a promise.EU-wide coverage with strong DACH community contributions; Austrian and German retail barcodes are well represented, and the nightly database export is small enough to run fully offline for Vienna grocery shopping.
CodeCheck food, cosmeticsproprietaryOwn scientific department plus partner data: EU CosIng, the California Department of Public Health, the Deutscher Allergie- und Asthmabund (DAAB, German Allergy and Asthma Association), the Hamburg consumer centre (VZHH), Greenpeace Switzerland and WWF.Free barcode and text-search scanning; the app stores list this as a free download and app-store privacy labels are the only documented source, so account requirements and server-side data handling are not independently confirmed here.Explicitly listed as available in Germany, Austria, Switzerland, the UK, the US and the Netherlands, a Swiss-German-founded (ETH Zurich spin-off) service with direct Austrian coverage.
ToxFox (BUND) cosmetics, toys, furniture, textiles, footwear, electronicsproprietaryThe EU REACH Regulation's Article 33 substances-of-very-high-concern (SVHC) list, built out through the EU LIFE project AskREACH; a barcode scan that finds no match triggers a direct question to the manufacturer, who owes a REACH-mandated answer within 45 days.Free app published by the Bund für Umwelt und Naturschutz Deutschland (BUND, German nature and environment federation); BUND's own GDPR notice for the app is public, covering what personal data it processes and why.Listed directly on the Austrian Apple App Store and co-promoted by the Austrian environmental NGO GLOBAL 2000, so Austria is an explicit, not just incidental, market for it, though its product catalogue leans toward cosmetics and household goods rather than groceries.
Fig food, allergens, dietsproprietaryOwn ingredient-tagging database covering 2,500-plus allergen, diet and food-dye tags across 300,000-plus products from more than 100 US grocery chains.Free to start, with 5 scans and unlimited searches per month on the free tier; account creation is not stated as required to begin scanning.US-only today; the vendor names the UK, Canada, Australia and New Zealand as planned next markets, with no continental-Europe timeline, so Fig is not usable for Austrian grocery shelves.
EWG's Healthy Living food, cosmetics, household cleaningproprietaryThe Environmental Working Group's own Skin Deep (cosmetics), Food Scores and Guide to Healthy Cleaning databases, combined to cover roughly 200,000 food, personal-care and cleaning products.Free to download, but account creation is part of onboarding; a paid premium tier adds further tools on top of the free core.The Environmental Working Group is a US nonprofit and its Food Scores database skews toward US retail assortments, so coverage of an Austrian supermarket basket is expected to be thin even though the app itself is globally downloadable.
HowGood food sustainabilityproprietaryOwn Latis ingredient-sustainability database, described as covering 250,000-plus products and over 1 million UPC codes across 60-plus sustainability categories (labor conditions, pesticide use, emissions and more).HowGood launched a public consumer barcode-scanning app in 2014; the company has since built out Latis as a B2B enterprise sustainability-intelligence platform sold to food brands and retailers, and current sources describe Latis, not a maintained consumer scanner, so consumer-facing scanning and any related privacy terms could not be confirmed as of 2026.Enterprise-facing today (adopted by large US food companies per press coverage) with no confirmed Austrian retail integration or consumer access route.
Scan Halal halal status, ingredient sourcingproprietaryOwn ingredient-sourcing database of 2 million-plus products, breaking questionable ingredients down by source (plant, animal, synthetic) rather than only returning a halal or haram verdict.Account requirements and server-side data handling are not detailed on the vendor's own pages or app-store listing found in this research; flagged here as unverified rather than assumed.Positioned for the US Muslim community market in its own marketing copy; Austrian retail coverage and EU halal-certification alignment are not confirmed in the sources found.
Yumi foodOSS GPL-3.0Reads directly from the Open Food Facts database; no separate proprietary product catalogue of its own.Its own README states the design goal directly: no account required, no tracking, and no data sent to third-party servers for profiling, with the health-score calculation run on-device rather than server-side.Inherits Open Food Facts' EU and Austrian barcode coverage since it is a thin client over that database; a small independent Flutter project, so maturity and maintenance continuity are a real caveat versus the larger apps on this list.

Yuka

proprietary

Analyzes: food, cosmetics

Data source: Own proprietary product database, seeded from Open Food Facts and Open Beauty Facts in 2017 to 2018, then layered with a proprietary contribution-verification system since January 2018; around 700,000 to 800,000 products as of December 2019.

Privacy: Free to scan with no account required for first use, but server-side handling of scan activity is not independently verified here: Yuka's own privacy policy documents collecting account details (name, email, password), device data (device ID, model, manufacturer, OS and version), IP address and, only with explicit permission, location; it does not state whether product analysis runs on-device or server-side, and it does not support the narrower claim that only anonymised lookups are sent to Yuka's servers.

Austria fit: Confirmed live in Germany and Switzerland (DACH neighbours) since 2021, but Austria was not on its own country launch list as of early 2020; since it draws on the same EU-wide barcode data as Open Food Facts, scanning products bought in Austria likely still resolves, just without an official Austrian rollout.

Open Food Facts

OSS

AGPL-3.0 (server and mobile app code); Open Database License / Database Contents License (the product data itself)

Analyzes: food

Data source: Own collaborative, crowdsourced database (Product Opener server plus the Android, iOS and web apps), with nightly full-database dumps and exports published for offline use.

Privacy: Scanning and browsing require no account; contributing edits or photos does. The server, app and database dumps are all self-hostable, so a fully offline, no-telemetry deployment is a real option, not just a promise.

Austria fit: EU-wide coverage with strong DACH community contributions; Austrian and German retail barcodes are well represented, and the nightly database export is small enough to run fully offline for Vienna grocery shopping.

CodeCheck

proprietary

Analyzes: food, cosmetics

Data source: Own scientific department plus partner data: EU CosIng, the California Department of Public Health, the Deutscher Allergie- und Asthmabund (DAAB, German Allergy and Asthma Association), the Hamburg consumer centre (VZHH), Greenpeace Switzerland and WWF.

Privacy: Free barcode and text-search scanning; the app stores list this as a free download and app-store privacy labels are the only documented source, so account requirements and server-side data handling are not independently confirmed here.

Austria fit: Explicitly listed as available in Germany, Austria, Switzerland, the UK, the US and the Netherlands, a Swiss-German-founded (ETH Zurich spin-off) service with direct Austrian coverage.

ToxFox (BUND)

proprietary

Analyzes: cosmetics, toys, furniture, textiles, footwear, electronics

Data source: The EU REACH Regulation's Article 33 substances-of-very-high-concern (SVHC) list, built out through the EU LIFE project AskREACH; a barcode scan that finds no match triggers a direct question to the manufacturer, who owes a REACH-mandated answer within 45 days.

Privacy: Free app published by the Bund für Umwelt und Naturschutz Deutschland (BUND, German nature and environment federation); BUND's own GDPR notice for the app is public, covering what personal data it processes and why.

Austria fit: Listed directly on the Austrian Apple App Store and co-promoted by the Austrian environmental NGO GLOBAL 2000, so Austria is an explicit, not just incidental, market for it, though its product catalogue leans toward cosmetics and household goods rather than groceries.

Fig

proprietary

Analyzes: food, allergens, diets

Data source: Own ingredient-tagging database covering 2,500-plus allergen, diet and food-dye tags across 300,000-plus products from more than 100 US grocery chains.

Privacy: Free to start, with 5 scans and unlimited searches per month on the free tier; account creation is not stated as required to begin scanning.

Austria fit: US-only today; the vendor names the UK, Canada, Australia and New Zealand as planned next markets, with no continental-Europe timeline, so Fig is not usable for Austrian grocery shelves.

EWG's Healthy Living

proprietary

Analyzes: food, cosmetics, household cleaning

Data source: The Environmental Working Group's own Skin Deep (cosmetics), Food Scores and Guide to Healthy Cleaning databases, combined to cover roughly 200,000 food, personal-care and cleaning products.

Privacy: Free to download, but account creation is part of onboarding; a paid premium tier adds further tools on top of the free core.

Austria fit: The Environmental Working Group is a US nonprofit and its Food Scores database skews toward US retail assortments, so coverage of an Austrian supermarket basket is expected to be thin even though the app itself is globally downloadable.

HowGood

proprietary

Analyzes: food sustainability

Data source: Own Latis ingredient-sustainability database, described as covering 250,000-plus products and over 1 million UPC codes across 60-plus sustainability categories (labor conditions, pesticide use, emissions and more).

Privacy: HowGood launched a public consumer barcode-scanning app in 2014; the company has since built out Latis as a B2B enterprise sustainability-intelligence platform sold to food brands and retailers, and current sources describe Latis, not a maintained consumer scanner, so consumer-facing scanning and any related privacy terms could not be confirmed as of 2026.

Austria fit: Enterprise-facing today (adopted by large US food companies per press coverage) with no confirmed Austrian retail integration or consumer access route.

Scan Halal

proprietary

Analyzes: halal status, ingredient sourcing

Data source: Own ingredient-sourcing database of 2 million-plus products, breaking questionable ingredients down by source (plant, animal, synthetic) rather than only returning a halal or haram verdict.

Privacy: Account requirements and server-side data handling are not detailed on the vendor's own pages or app-store listing found in this research; flagged here as unverified rather than assumed.

Austria fit: Positioned for the US Muslim community market in its own marketing copy; Austrian retail coverage and EU halal-certification alignment are not confirmed in the sources found.

Yumi

OSS

GPL-3.0

Analyzes: food

Data source: Reads directly from the Open Food Facts database; no separate proprietary product catalogue of its own.

Privacy: Its own README states the design goal directly: no account required, no tracking, and no data sent to third-party servers for profiling, with the health-score calculation run on-device rather than server-side.

Austria fit: Inherits Open Food Facts' EU and Austrian barcode coverage since it is a thin client over that database; a small independent Flutter project, so maturity and maintenance continuity are a real caveat versus the larger apps on this list.

The onboarding flow this needs

None of the services above ask for values, a diet profile or an allergen list before they let someone scan. A store-shopping scanner built here would flip that order, front-loading the questions that make a single scan meaningful and deferring the one permission that is hardest to take back:

  1. Import values weights. Pull the same ownership, regional, seasonal, organic, packaging and fair-labor weights a person already set in the ideal-groceries tool, so a scan in the store scores against the values a person already chose rather than a generic default.
  2. Load the diet profile. Read the profile already stored for this repo's shared contract, via loadDietProfile from src/lib/shared/diet_profile.ts (style, exclusions, notes). This research page does not import it; a real scanner build would.
  3. Collect the allergen list. A short, explicit list beyond what the diet profile's exclusions already covers, since an allergen is a hard stop, not a preference.
  4. Set a strictness slider. One scale from "flag only the worst matches" to "flag anything off-profile", so the same scan can whisper or shout depending on how the person actually wants to shop.
  5. Ask for camera permission last, not first. By the time the app needs the camera, it already knows enough to make the scan useful; asking earlier just trains people to tap "allow" on autopilot before they understand why.
  6. Keep every step local. Values weights, diet profile, allergen list, strictness setting and scan history all stay on-device (the same local-storage pattern diet_profile.ts already uses); nothing above requires an account or a server round trip to configure.

Why the privacy full stack is semi-required

A camera pointed at shelves for an entire shopping trip is not one photo, it is continuous personal data: what a person buys, how often, at which store, on which day. Scoring that against a values profile and a diet profile makes it more sensitive, not less, since the output ties a health condition or an ethical stance to a real-time location. The concept this project's quest framing keeps returning to is that this kind of analysis layer belongs on hardware, software and infrastructure that is per default not surveilled and out of reach of AI models, built on a web of trust and private devices rather than a vendor's cloud.

That is a demanding bar for a phone app, and most of the field above cannot meet it: seven of 9 services here are proprietary, with the phone-to-server round trip and the retained account state that implies. The practical near-term path is Open Food Facts' own offline, self-hostable database (AGPL-3.0 code, ODbL data, openfoodfacts-server on GitHub), which exports nightly and can run entirely on-device or on hardware a person controls, with no query ever leaving it. Yumi, the small open source client already built against that same database, is the existence proof that a no-account, on-device scanner is buildable today, not just theoretically possible.

Honest gap

Nothing here is a working scanner. The actual point-a-camera-and-get-a-verdict app, wired to local values weights, a diet profile and an allergen list, is a future wave. This page is the researched foundation that wave would start from: the service landscape, the onboarding order, and the privacy stack the analysis layer would need to actually earn the trust it is asking for.