Calibras digitalizes manufacturing workflows and provides strategic digital transformation consultancy for manufacturing companies.
We proud ourselves in our manufacturing roots which we merge with full-stack software expertise, delivering our clients with in-depth insights and tailored solutions that provide tangible value.
By combining our multi-disciplinary capacities, we develop and deploy production grade multimodal manufacturing AI agents, domain specific data science & ML pipelines, 3D geometry processing solutions, computer vision measurements and process control systems, all integrated into your factory or laboratory framework, ERP and MES system.
We live in a manufactured world, and physical value-added remains the foundation of all industrial economies and strategic relevance. Modern manufacturing is digital. This is why it is imperative to have a reliable partner on your side, that can cut through the noise and provide guidance at the overlap of physical and digital world.
With discernment and loyalty, we stand ready to be just that.
3D scanning workflow automation for an industry leading 3D printing bureau: Parts on Demand
* Automated 3D scanning with generated dashboards.
Developed and deployed an intelligent hardware-software module for Techvolver‘s pipette calibration robotics system.
* OCR + mechatronics lab automation module.
Calibrated Computer Vision 2D dimensional measurement system. Dedicated thresholding, edge detection and oriented bounding box based algorithm had been developed and calibrated against tactile measurement results. Optical setup included IDS industrial grade camera, dedicated ring light and polarization filter. Achievable resolution: 0.05 mm; Achievable accuracy: 0.25 mm
* Dimensional measurement setup & example results.
Showcase of our industrial computer vision solutions developed with the support of DL Industrial Automation AB, using Impact software. Pipelines are highly adaptable to various manufacturing use cases, eg. defect detection, measurements, AI based OCR etc. End-to-end setup was using robust industrial grade hardware – P30M 100-000 ML Industrial Smart Camera.
* Industrial camera and example output from developed solutions.
Development and maintenance of cloud-to-machine layer for a leading dental MOS provider: Oqcam
* IoT & gateways architecture for a dental lab
Project delivered 2 automated Capability Sixpack pipelines, capable of clustering the input dataset into smaller subsets and performing capability analysis on each. Input dataset could be acquired either from external input or pulled from Minitab. Capability analysis could be performed and results can be plotted using either a custom pipeline or using Minitab’s functionality. Output can be placed into a dedicated folder structure and visualized in Minitab.
* Resulting Capability Analysis Sixpack – sample from custom CLI pipeline
Project scope was focused on validation of a new Werth TomoScope XS CT for the use in production. This was achieved via a series of step gauge measurements, calibrated with with Zeiss CMM with attached uncertainty estimation. During the execution of the project, an academic investigation on the effect of surface roughness, mesh filtering and outlier removal was done, resulting in a custom software package developed for the purpose of data ingest and visualization, as well as a custom procedure for estimation of the effect of certain factors (based on variability analysis).
* CMM physical filtering, step gauge & CMM measurements, CT measurements.
Fully integrated AI based process monitoring solution, accurately predicting a critical process parameter from the thermal camera and webcam input.
The solution was developed for Additive Manufacturing (AM) process, with the task of predicting printed part warpage. System runs on a single chip computer, and can acquire and process data from 2 sensors, perform complex image processing on the input data before feeding it into a AI model, gather predictions, and finally, render results in a developed GUI (which runs in a local server app).
* Project workflow.
Project developed a custom metrological ML software framework, capable of generating artificial point cloud measurements from confocal microscope data. Project involved manufacture of a custom measurement object, as well as measurement acquisition on Olympus Lext machine. A custom Variational Autoencoder (VAE) model with convolutional (CNN) encoder block was developed to perform reconstruction. Measurand was constructed as an ISO 2-point distance and finally, comparison of measurement results between real and generated data was given.
* Reconstructed confocal microscope step-height measurement.
Generation of Point Cloud measurements using a Stereo Camera. A custom continuous acquisition method and ICP based reconstruction pipeline (with different loss functions) had been developed, enabling full scene and/or measurement object reconstruction. Raw input point clouds, generated via disparity map from the camera, had been continuously accumulated, filtered for outliers, downsampled and finally, overlapped using an ICP algorithms with various loss functions (i.e. Tukey loss).
* Raw stereo camera single frame acquisition – disparity map based point cloud.
Project applied advanced Optical Character Recognition (OCR) ML models to extract critical information from mechanical CAD drawings. A separated image processing pipeline was implemented to extrapolate relevant ROIs as the input for the OCR model and group the drawing into separate components (Infoblock frame, GPS dimensions & tolerances, drawing frame etc.).
* Final output of the OCR pipeline.
Main project target was to develop a ML framework, capable of accurately predicting cutting forces in X and Y direction, from a 2D CNC machining simulation data. Extensive Data Mining system had been established, capable of parsing and merging complex simulation output into a usable dataset. Data was subjected to outlier removal (IQR and Z-score) and BoxCox transformations, before being fed into a custom MLP model. Final results achieved MAE of 0.8195 % when comparing to average load in X direction and 4.0187 % in Y direction.
* Deform 2D CNC simulation + resultant load prediction.
3D scanning workflow automation for an industry leading 3D printing bureau: Parts on Demand
* Automated 3D scanning with generated dashboards.
Developed and deployed an intelligent hardware-software module for Techvolver‘s pipette calibration robotics system.
* OCR + mechatronics lab automation module.
Calibrated Computer Vision 2D dimensional measurement system. Dedicated thresholding, edge detection and oriented bounding box based algorithm had been developed and calibrated against tactile measurement results. Optical setup included IDS industrial grade camera, dedicated ring light and polarization filter. Achievable resolution: 0.05 mm; Achievable accuracy: 0.25 mm
* Dimensional measurement setup & example results.
Showcase of our industrial computer vision solutions developed with the support of DL Industrial Automation AB, using Impact software. Pipelines are highly adaptable to various manufacturing use cases, eg. defect detection, measurements, AI based OCR etc. End-to-end setup was using robust industrial grade hardware – P30M 100-000 ML Industrial Smart Camera.
* Industrial camera and example output from developed solutions.
Development and maintenance of cloud-to-machine layer for a leading dental MOS provider: Oqcam
* IoT & gateways architecture for a dental lab
Project delivered 2 automated Capability Sixpack pipelines, capable of clustering the input dataset into smaller subsets and performing capability analysis on each. Input dataset could be acquired either from external input or pulled from Minitab. Capability analysis could be performed and results can be plotted using either a custom pipeline or using Minitab’s functionality. Output can be placed into a dedicated folder structure and visualized in Minitab.
* Resulting Capability Analysis Sixpack – sample from custom CLI pipeline
Project scope was focused on validation of a new Werth TomoScope XS CT for the use in production. This was achieved via a series of step gauge measurements, calibrated with with Zeiss CMM with attached uncertainty estimation. During the execution of the project, an academic investigation on the effect of surface roughness, mesh filtering and outlier removal was done, resulting in a custom software package developed for the purpose of data ingest and visualization, as well as a custom procedure for estimation of the effect of certain factors (based on variability analysis).
* CMM physical filtering, step gauge & CMM measurements, CT measurements.
Fully integrated AI based process monitoring solution, accurately predicting a critical process parameter from the thermal camera and webcam input.
The solution was developed for Additive Manufacturing (AM) process, with the task of predicting printed part warpage. System runs on a single chip computer, and can acquire and process data from 2 sensors, perform complex image processing on the input data before feeding it into a AI model, gather predictions, and finally, render results in a developed GUI (which runs in a local server app).
* Project workflow.
Project developed a custom metrological ML software framework, capable of generating artificial point cloud measurements from confocal microscope data. Project involved manufacture of a custom measurement object, as well as measurement acquisition on Olympus Lext machine. A custom Variational Autoencoder (VAE) model with convolutional (CNN) encoder block was developed to perform reconstruction. Measurand was constructed as an ISO 2-point distance and finally, comparison of measurement results between real and generated data was given.
* Reconstructed confocal microscope step-height measurement.
Generation of Point Cloud measurements using a Stereo Camera. A custom continuous acquisition method and ICP based reconstruction pipeline (with different loss functions) had been developed, enabling full scene and/or measurement object reconstruction. Raw input point clouds, generated via disparity map from the camera, had been continuously accumulated, filtered for outliers, downsampled and finally, overlapped using an ICP algorithms with various loss functions (i.e. Tukey loss).
* Raw stereo camera single frame acquisition – disparity map based point cloud.
Project applied advanced Optical Character Recognition (OCR) ML models to extract critical information from mechanical CAD drawings. A separated image processing pipeline was implemented to extrapolate relevant ROIs as the input for the OCR model and group the drawing into separate components (Infoblock frame, GPS dimensions & tolerances, drawing frame etc.).
* Final output of the OCR pipeline.
Main project target was to develop a ML framework, capable of accurately predicting cutting forces in X and Y direction, from a 2D CNC machining simulation data. Extensive Data Mining system had been established, capable of parsing and merging complex simulation output into a usable dataset. Data was subjected to outlier removal (IQR and Z-score) and BoxCox transformations, before being fed into a custom MLP model. Final results achieved MAE of 0.8195 % when comparing to average load in X direction and 4.0187 % in Y direction.
* Deform 2D CNC simulation+ resultant load prediction.
“Working with Calibras has been a true game-changer for our quality assurance. Their expertise in automated 3D scanning and metrology has transformed our QC process. We’ve not only significantly reduced the time it takes to perform quality checks, but the data integration into our dashboards gives us unprecedented insight into production trends. This allows us to proactively manage and tighten our tolerances, ensuring we consistently deliver on our promise of setting new standards in part quality and reproducibility. Calibras is a vital partner in our mission to set the new standard in industrial 3D printing.”
CEO & Geek-in-chief
“We’ve been very satisfied with Calibras’ work on our loT stack. They effectively managed the cloud-to-machine layer for our 3D printers, supported multiple loT protocols, and handled service orchestration both in the cloud and on local computers. Their maintenance of our gateway fleet and custom integrations, such as workflow automation and SMB network drives, secured a smooth transition between our legacy and new loT solution. Calibras also provided solid and highly responsive support directly to our customers and always represented our company in the most professional and reliable manner.”
Head of Development
“Calibras’ work on our lab automation solution proved to be an extremely valuable addition to our system. They developed a very complex tailored module for us, combining OCR technology and advanced mechatronics, enabling our system to have reproducible and reliable volume information, as well as capability to adjust it. Their hardware – software expertise truly makes them capable of delivering value to us and our pharma solutions. We consider them an important partner in our product development journey.”
Co-Founder & CEO
© 2026 Calibras. All rights reserved.
