an engineer to clean data.
in one session, not days of scripting.
Data quality checks
Every issue. Exactly where it happens.
Automated checks flag every issue on the signal. Preview a fix against the raw data before you apply it.
Transparency
Not a black box. You stay in control.
Rule and threshold for every check
See exactly why a signal was flagged.
Preview before you apply
Compare the fixed signal with the raw one first.
Raw data never overwritten
Fixes create a cleaned version. The original stays.
Nothing changes without your approval
You choose which issues to fix and how.
Events & labeling
Label any event.
Built-in today, your own soon.
Built-in labels for maneuvers and driving context run automatically.
Coming soon
Need an event the built-in labels don’t cover? Describe it in your own words –
AI turns it into a custom label you check on your data before saving.
1 Describe
Type the event in your own words, e.g. “Label when vehicle speed stays above 120 km/h for more than 10 seconds”.
2 Review the rule
See the proposed rule in plain language and the events it finds in your data.
3 Save
Add the label to your library. It runs on every future dataset.
Data & deployment
Works with your data. Runs where it has to.
EU cloud
TISAX certified
Your cloud
Runs in your own VPC
On-premise
Data never leaves your network
Pricing
Start small. Scale when you need to.
DataPilot S
For a first team working with vehicle data.
DataPilot M
For several teams and larger test campaigns.
DataPilot Enterprise
Tailored to your users, data volume and deployment needs.
Every plan starts with a free 30-day trial.
FAQ
Yes. DataPilot runs the quality checks, proposes fixes and labels driving events automatically. Your engineers review the results instead of writing scripts.
DataPilot matches your signal names to standard DataPilot signals automatically. You review each match and can change it.
You don’t have to. Every flagged issue shows what was detected and why it was flagged. Before any fix is applied, you see a preview of the corrected signal alongside the original. Nothing changes in your data without your explicit confirmation, and your raw files are never touched – cleaned output is always a separate version.
You don’t need to. Each built-in label is a rule with fixed thresholds – for example, hard braking starts when longitudinal acceleration drops below −0.7 g for at least 0.5 s. DataPilot finds the start and end of every event, and you review them per trip before export.
On-premise and data-lake deployment is available for enterprise customers. Book a demo to discuss deployment options.
DataPilot is built for the engineers who work directly with vehicle data – data scientists, test engineers, calibration engineers, and ML engineers who spend time writing scripts to clean and validate datasets before analysis.
You can use DataPilot free for 30 days. When the trial ends, a paid subscription is required to continue using the platform.
DataPilot starts with a free 30-day trial. After that, you choose an annual package – S, M or Enterprise. See plans and prices
MAT, CSV and Parquet. MF4 support is coming soon.
The cleaned and labeled dataset – in the format you uploaded or as Parquet. Labels come as extra columns next to your signals, so your existing scripts read them as they are. The quality report exports as CSV. Your raw files stay untouched.
We are preparing a sandbox with a sample dataset – no setup, no upload. Until it opens, start a 30-day trial with your own data, or book a demo and we’ll walk you through the product on ours.
The score (0–100) is the share of valid samples across all signals in your dataset. Every flagged issue shows the check that found it, so you can see what lowered the score.
Building the next generation of automotive data infrastructure. Certified TISAX | EU-Hosted | © 2026 COMPREDICT GmbH – Darmstadt, Germany
All Rights Reserved
