Skip to content

AI camera system for e-waste recycling · prototype in development

AI-based sampling of e-scrap: cost effective, fast and secure.

metora is an innovative deep tech start up developing an AI based camera system for sampling of e-scrap. This camera system can be easily installed on the conveyor belt and determines the precious and base metal content of the passing e-scrap – by orders of magnitude more cost effective, faster and more secure than conventional sampling methods.

Rendering: blue camera housing with the metora logo on a black gantry above a conveyor belt
Prototype design (rendering)

Benefits

  • High Accuracy

    Cutting edge AI model matches accuracy of conventional sampling methods.

  • Cost Effective

    Substantially more cost effective than conventional sampling methods.

  • Secure

    Inherently immune to fraud and human errors.

  • Fast Results

    Results delivered in a matter of hours.

How metora works

  1. Detect

    The camera detects circuit boards passing beneath it, tracks them and selects a usable capture for each.

  2. Determine

    An AI model trained on metora's own image and laboratory data determines the board's metal content from that capture — gold, silver, palladium, platinum, copper and aluminium.

  3. Trace

    Captured boards land in the web app with image, ID and estimate — groupable into lots, with totals and distributions per lot.

From belt to estimate

Illustrative demo · example data
Selected example captureA-01
Boardsper board

Example sequence overview

Results: 3 of 3

Total gold0.090 gSum of 3 of 3 values
Total silver0.120 gSum of 3 of 3 values
Total palladium0.012 gSum of 3 of 3 values
Total platinum0.006 gSum of 3 of 3 values
Total copper54.000 gSum of 3 of 3 values
Total aluminium18.000 gSum of 3 of 3 values
What the demo shows — and what it doesn't

The demo runs entirely in your browser with fixed example data: three boards are detected, captured and given example estimates for Au, Ag, Pd, Pt, Cu and Al. It shows no real camera capture, no model output and no connection to a device. The steps detect, capture, estimate and overview follow the system's flow; values and timing are illustrative.

Real results live in the web app (external link, login required)

Technology and current stage

The model learns from photographs of circuit-board fragments together with their actual material composition. The composition of the training data was determined in a purpose-built analysis workflow.

The current prototype was developed for several categories of circuit boards and tested under real conditions on a conveyor. As the next step we are looking for pilot and industrial partners to validate the system on their material and in their workflows, side by side with existing sampling.

Every captured board as a record: image, ID, estimates, lot.

The metora web app brings all estimates together: per captured board with image, ID, time stamp and an estimate per metal, per lot as totals and distribution. Access is intended for customers and pilot partners.

Example lot (synthetic)

3 Boards
Demo IDper board
A-01
  • Gold (Au)10 mg
  • Silver (Ag)20 mg
  • Palladium (Pd)2 mg
  • Platinum (Pt)1 mg
  • Copper (Cu)12.000 g
  • Aluminium (Al)4.000 g
B-02
  • Gold (Au)30 mg
  • Silver (Ag)40 mg
  • Palladium (Pd)4 mg
  • Platinum (Pt)2 mg
  • Copper (Cu)18.000 g
  • Aluminium (Al)6.000 g
C-03
  • Gold (Au)50 mg
  • Silver (Ag)60 mg
  • Palladium (Pd)6 mg
  • Platinum (Pt)3 mg
  • Copper (Cu)24.000 g
  • Aluminium (Al)8.000 g
Total gold0.090 g
Total silver0.120 g
Total palladium0.012 g
Total platinum0.006 g
Total copper54.000 g
Total aluminium18.000 g

Who we build for

Recyclers and refiners

See the estimated precious and base metal content of incoming circuit boards while the material is moving — documented per captured board, summarised per lot — traceable in the web app.

Industrial partners

The system mounts above an existing conveyor. We are looking for partners in plant engineering, conveying technology and analytics to validate the prototype with us on real lines.

Investors

metora combines its own hardware, an AI model and analysis software. The prototype has been tested; we are now looking for pilot and industrial partners. Supported by aws, FFG, Land Tirol, greenstart and Startup.Tirol.

metora in Japan

metora is a J-Bridge partner of JETRO (Japan External Trade Organization) and listed in its 2026 J-Bridge database of promising overseas companies. J-Bridge is JETRO's open-innovation platform connecting Japanese companies with start-ups from abroad. Japanese companies can request a meetup through J-Bridge or contact us directly.

News

Events, awards and partnerships around metora.

All news
Laurenz Birnbaum and another person holding the wooden greenstart star reading “greenstart 2025 metora”
Laurenz Birnbaum (metora) at the greenstart kick-off 2025 with the award for metora.Photo: © Niklas Stadler / Klima- und Energiefonds

Programme14 May 2025

metora among the TOP 10 of greenstart 2025

The Austrian Climate and Energy Fund has selected metora for the TOP 10 of the 2025 anniversary cohort of its start-up programme greenstart.

UpdateSource as of 1 April 2026

metora is a J-Bridge partner of JETRO

metora is a partner of the J-Bridge platform of JETRO (Japan External Trade Organization) and listed in its 2026 J-Bridge database of promising overseas companies — the direct route to metora for Japanese companies.

Read more

Company

Developed in Innsbruck

Our founders

Portrait of Laurenz Birnbaum

Laurenz Birnbaum

Co-founder – optical systems and e-waste sampling

M.Sc. Mechatronics, M.Sc. Engineering Physics
Portrait of Aaron Schulz

Aaron Schulz

Co-founder – AI, computer vision and software

M.Sc. Physics, B.Sc. Computer Science

Contact

Contact us

Email

Your choice pre-fills the email subject.

Please tell us briefly who you are and what you have in mind — a pilot, a partnership or an investment.