# Errors Below is a table of the main error codes and checks that may occur during KYC procedures on the NVGlobal platform. For convenience, the table includes a concise description, a user-facing comment, and a brief explanation of the technology behind each check. The reference is aligned with the internal “Errors” section and example responses of `/kyc/process` and `/kyc/session/status`. | No. | Check Code | Description | User Comment | Technology Description | | --- | ----------------------------------------- | --------------------------------------------- | ------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ | | 1 | \[general] bad image | The uploaded file is not an image | You uploaded a file that is not an image | The system validates the file type. Only JPEG and PNG formats are supported. If the file does not match these formats, an error is returned. | | 2 | \[general] too small image | The image is too small for proper recognition | The image resolution is too low; please upload a higher‑quality photo | The system checks the image resolution in pixels. If below the required threshold, an error is returned recommending a higher-quality upload. | | 3 | \[general] no documents on image | No documents detected on the image | Make sure you are photographing the required document | A neural network detector analyzes the image. If no objects resembling documents are found, an error is returned. | | 4 | \[general] not a selfie | No face detected outside the document | Clean the camera and try again, ensure the image is in focus | A face detector checks for a face outside the detected document area. If no face is found, the system requests a clearer image. | | 5 | \[general] low light | Insufficient lighting | Use better lighting or move to a brighter location | The system evaluates the illumination level. If below threshold, an error is returned. | | 6 | \[general] age requirement mismatch | Age does not meet requirements | Your age does not meet the requirements | A neural network estimates age with ±4 years accuracy. If it falls below the threshold or does not match the document-based age, an error is returned. | | 7 | \[general] black and white image | Black-and-white image | Do not use filters — the photo must be in color | The system analyzes color channels to detect grayscale or desaturated images. | | 8 | \[general] too many faces | Multiple faces detected on document image | Ensure only the document is in the frame | The system uses a face detector; if more than one face is found, an error is returned. | | 9 | \[document] unknown document type | Unknown document type | Ensure you are photographing the required document | The system classifies the document by visual features. | | 10 | \[document] bad document type | Incorrect document type | Ensure you are photographing the required document | The system identifies document type via a dedicated detector; mismatches trigger an error. | | 11 | \[document] too small image | Document is too small | Move the camera closer | The detector checks document size relative to frame; if too small, an error is returned. | | 12 | \[document] bad pages | Wrong document page(s) | Ensure you are photographing the required pages | Page classifiers verify correct side/spread of the document. | | 13 | \[document] no face | No face found on document | Ensure the portrait is visible | Face detector could not find a portrait zone on the document. | | 14 | \[document] text fields not visible | Fields not visible | Ensure the image is sharp and fields are not covered | OCR and segmentor detect field regions; if low visibility, an error is returned. | | 15 | \[document] bad mrz checksum | MRZ checksum error | Ensure the image is clear | OCR reads MRZ lines and validates checksums; mismatch triggers error. | | 16 | \[fraud] display or screen | Screen/monitor photo | A photo of a screen cannot be used; take a photo of the original document | Anti-fraud model detects moiré, subpixel patterns, and screen reflections. | | 17 | \[fraud] image edited | Image edited | Edited photos cannot be used | ML classifier detects signs of editing (inpainting, smoothing, metadata anomalies). | | 18 | \[fraud] image printed | Printed image | Do not use printed copies | Classifier detects paper texture and raster artifacts. | | 19 | \[face] bad face matching | Faces do not match | Make a selfie with the document clearly visible | The system compares document and selfie embeddings; low score triggers error. | | 20 | \[face] not all faces found | Not all faces detected | Ensure the image is clear | Liveness requires several images; missing detections trigger error. | | 21 | \[fraud] black list | Blacklist match | You did not pass the verification | Face embedding matches an internal watchlist. | | 22 | \[legal] legal proceeding existence | Legal proceedings found | You did not pass the verification | External database indicates open legal cases. | | 23 | \[legal] document not valid | Document invalid | Your document is invalid | Validation via structure, MRZ, dates, and external registries. | | 24 | \[system] balance too low | Insufficient balance | Top up your balance | Billing system reports insufficient funds. | | 25 | \[system] exception | Internal system error | Try again | Unhandled backend exception; session fails. | | 26 | \[document] bad quality | Poor document quality | Ensure the image is sharp and clean | Quality metrics: sharpness, noise, exposure, glare. | | 27 | \[general] too many tries | Too many attempts | You exceeded the allowed number of attempts | Anti-abuse limit on attempts. | | 28 | \[fraud] manipulation detected | Manipulation detected | Take the photo again | Aggregated anti-fraud score triggers rejection. | | 29 | \[fraud] document modified | Document modified | Document must not contain physical alterations | Detector identifies inserts, overlays, texture mismatches. | | 30 | \[fraud] logical mismatch | Logical mismatch | You did not pass verification | Decision Engine finds inconsistent dates, series, numbers. | | 31 | \[fraud] bad document photo | Bad document portrait | Do not cover the portrait | Portrait region shows occlusions or artifacts. | | 32 | \[fraud] mobile device | Mobile device screen photo | Do not use screen photos | Mobile-screen classifier detects subpixel and reflection patterns. | | 33 | \[fraud] document damaged | Document damaged | The document is damaged | Anti-fraud detects tears, wear, deformation. | | 34 | \[document] cropped | Document cropped | Document must be fully visible | Boundary detector identifies incomplete contour. | | 35 | \[document] no signature | Signature missing | No signature in the document | Signature zone segmentation detects absence of handwritten stroke. | | 36 | \[document] no stamp | Stamp missing | No stamp present | Stamp detector checks for circular/seal patterns. | | 37 | \[document] no required objects | Missing required elements | Required elements missing | Required fields: photo, signature, stamp, MRZ, etc. | | 38 | \[document] text field has low confidence | Low field confidence | Field recognized with low confidence | OCR marks field confidence as low. | | 39 | \[fraud] multiple documents | Multiple documents detected | Different documents detected within one session | Anti-fraud rule triggered by inconsistent document instances. | | 40 | \[document] expired | Document expired | Document validity period has ended | Validity date check via OCR/external DB. | | 41 | \[system] service timeout | Service timeout | The service did not respond in time | SLA timeout reached. | | 42 | \[document] other error | Other document error | Document does not meet requirements | Generic category for non-standard mismatches. | | 43 | \[face] is deepfake | Deepfake detected | Provide an original photo | Deepfake detection identifies GAN artifacts. | | 44 | \[face] is not alive | Liveness check failed | Provide a photo of a real, live person | Liveness engine detects absence of real presence (micro‑motions, series analysis). | *Note: These codes are also returned inside the `errors` arrays of KYC API responses, which allows mapping them to specific processing steps.*