Dashcam Camera Profiles

An open, community-maintained reference for dashcam card layouts, recording conventions, and privacy-preserving card scans.

This repository is useful whether or not you use Dashcam Offloader. Its goal is to make real, documented camera evidence reusable by any open tool.

Browse the reference

The static browser in docs/ makes the technical data readable without installing an app. It includes searchable camera pages for folder structures, filename rules, driving and parking behavior, measured codec/resolution/FPS/ bitrate data, manuals, evidence levels, and known gaps.

To preview it locally:

python3 -m http.server 8000 --directory docs

Then open http://localhost:8000. The browser has no account system, analytics, cookies, or runtime dependencies.

Methodology and evidence

Each profile keeps its evidence boundary visible. A card-validated claim is confirmed by a sanitized scan of a real card. A manual-backed claim comes from official documentation and still awaits card confirmation. A technical-sample records measured video facts without claiming that the card's folder or filename behavior is known. A catalog-hint identifies a model while making no layout-specific claim.

Measured data describes one recorded configuration, not a universal product specification. Resolution, frame rate, HDR, quality mode, firmware, region, and connected cameras can change a result. A product name alone is never enough to infer folder behavior or measured bitrate. The browser shows a recorded configuration and settings note when the evidence supplies them, and leaves those details unknown when it does not.

Contribute a card scan

You do not need to own or install Dashcam Offloader to contribute.

1. Read the privacy guide.
2. Create a JSON file that follows the card-scan schema.
3. Start from the example, remove anything identifying, and open an issue or pull request.

For now, non-technical contributors can also send a sanitized scan package to the Dashcam Offloader project through its in-app Learn Card workflow. A standalone scanner and browser upload flow are planned; neither is required to participate in this repository.

What belongs here

• Reviewed camera profiles, with their evidence level and sources.
• Sanitized, structural card scans that help validate those profiles.
• Official manual or product-page links that support documented behavior.
• Corrections to folder names, filename patterns, channels, recording modes, codecs, resolutions, and parking behavior.

Useful contributions include a generalized filename pattern such as REC_YYYYMMDD_HHMMSS_F.MP4, the exact case of an observed folder name, a manual link that documents a parking mode, or a local video summary that says which cameras were connected and which quality setting was selected. A video file itself is never needed.

What must never be submitted

• Video, photo, thumbnail, audio, GPS, or route data.
• Serial numbers, device IDs, MAC/Bluetooth IDs, Wi-Fi names or passwords, accounts, tokens, and full configuration dumps.
• Personal folder paths, volume labels, original filenames, license plates, or contacts in a public pull request or issue.

See CONTRIBUTING.md for the review process.

Evidence levels

• card-validated: a sanitized real-card scan has confirmed the behavior.
• manual-backed: official documentation supports it, but a card has not yet validated it.
• technical-sample: measured video facts are available, but folder or filename behavior remains unknown.
• catalog-hint: a model is known to exist, but layout-specific support is not claimed.

An evidence level is a boundary, not a quality score. A profile does not claim unobserved behavior simply because a similar model has it.

Repository layout

• schemas/ - stable public formats for scan and profile data.
• examples/ - synthetic examples safe to copy and modify.
• profiles/ - the canonical, reviewable technical profile for each camera.
• docs/ - the generated data index, searchable frontend, and privacy guide.
• scripts/ - deterministic import, validation, privacy, and index tooling.

Status

The repository currently contains 68 camera references: 56 camera profiles with card, manual, or catalog evidence, plus 12 references built from measured video metadata where the card layout is not yet known. Missing information is shown as missing rather than inferred from a related model.

Run python3 scripts/verify.py before submitting a profile change. It checks the canonical profiles, regenerated browser index, representative driving and parking data, filenames, folder layouts, manuals, and privacy boundaries.

License

Data in this repository is licensed under CC BY 4.0. Code or tools that consume it keep their own licenses.