Fixed costs for unlimited data
Monitor your LLM, ML model, data pipeline, data stream, or dataset - any volume, anywhere.
All plans include:
Monitoring for ML models, LLMs, data pipelines, or data streams
Monitoring 100% of the data, without sampling
Support for structured and unstructured data
Access to the full Observatory platform
6+ months of data retention
Privacy-preserving integration
Ability to sign up via the AWS Marketplace
STARTER
ML and Data Monitoring for individuals
Included:
- One user
- Up to two Projects
- Up to 200 features or columns/Project
- Five segments/Project
- 10M predictions/month
- Daily monitoring
- Community support
- No credit card required
Free
Up to 2 Projects
Get started for freeBOOK A DEMOEXPERT
For small and growing teams
Included:
- Up to five users
- One organization
- Up to 200 features or columns/Project
- Five segments/Project
- 100M predictions/month
- Daily or weekly monitoring
- Hourly monitoring for one Project
- Community and email support
- Additional projects available: up to 10
$125
/month
Up to 5 Projects
Buy nowENTERPRISE
For advanced scale, security and support needs
Included:
- Unlimited users
- Custom organizations
- Custom projects
- Custom features
- Custom segments
- Custom monitoring
- Unlimited predictions
- Custom data retention
- Real-time monitoring
- Enterprise support 12x5 or 24x7
- Private Slack channel
- Training, onboarding, and workshops
CUSTOM
Contact usSTARTER | EXPERT | ENTERPRISE | |
|---|---|---|---|
Free plan | $125/month Annual discount available | Custom pricing | |
Scale | |||
| Number of Projects | 2 Projects | 5 Projects | Custom Projects |
| Number of rows/predictions | 10M per month | 100M per month | Unlimited |
| Users | 1 user | Up to 5 users | Unlimited |
| Features or columns | 200 per Project | 200 per Project | Unlimited |
| Free test projects | |||
| 100% of the data monitored, no sampling | |||
Data | |||
| Data retention | 6 months | 6 months | Custom data retention |
| Real-time metrics | |||
| Streaming and batch data | |||
| Tabular, image, text, video, audio data | |||
| Data segments (prediction slices) | up to 5 per Project | up to 5 per Project | Custom |
| Privacy-preserving data profiling | |||
ML/AI Models | |||
| Model comparison | |||
| Training data baseline monitoring | |||
| All common model types | |||
| Large language models (LLM) | |||
| Ensemble models | |||
| Embeddings | |||
Monitors and alerts | |||
| Data quality monitors | |||
| Data drift monitors | |||
| Model performance monitors | |||
| Zero-configuration monitor set-up | |||
| Seasonal and custom monitors | |||
| Daily monitor granularity | |||
| Hourly monitor granularity | |||
| Custom monitor granularity | |||
| Real-time monitor granularity | |||
| Email and Slack notifications | |||
| PagerDuty and ServiceNow notifications | |||
| Custom Webhook actions | |||
| Retraining triggers | |||
| API access | |||
Platform Capabilities | |||
| Time machine view of data and model health | |||
| Interactive data investigation | |||
| Root cause analysis and debugging workflow | |||
| Model performance dashboard | |||
| Executive summary dashboard | |||
| Explainability-powered monitoring | |||
| Correlation of alerts across pipeline steps | |||
Security and Compliance | |||
| SOC 2 Certified | |||
| SAML SSO | |||
| Role-based access controls | |||
| Custom deployment model | |||
Support | |||
| Community Slack | |||
| Email support | |||
| 12x5 or 24x7 support availability | |||
| Dedicated Customer Success Data Scientist | |||
| Private Slack channel |
Frequently Asked Questions
Yes! There are two ways to use WhyLabs for free.
The WhyLabs Starter plan is free to use now—and forever—for two Projects. You get all of the convenience of monitoring for data drift, data quality, and model performance combined with continuous data logging and profiling for unlimited data volumes along with many other great features, all included for free.
Secondly, the WhyLabs Enterprise plan offers a free trial period that’s optimized for your unique requirements. Please contact us and we will be happy to work out a custom trial period.
Simply click here to create a free account and get started! The platform will guide you through all steps you need to onboard in just a few minutes.
For an overview of the process, you can take a look at the Onboarding guide in the documentation center.
Alternatively, contact us to schedule a call with our team. We’re happy to help you learn more.
Upgrading is easy when you’re ready to move up from the Starter plan!
When you want to start monitoring more Projects, simply add them to your Project dashboard, and we’ll contact you about your account upgrade.
To upgrade to Enterprise, please contact us
What is a Project?
A Project is any ML model, data pipeline, data stream, or dataset that you want to monitor in WhyLabs. The WhyLabs UI treats each of these items as a Project, so you can have flexibility in what you want to monitor.
What is a Feature?
A feature is a measurable property of the object you’re trying to analyze. In datasets, features appear as columns. In the WhyLabs UI, “Feature” is used interchangeably with “column”.
What is a Segment?
A Segment is a subset of your data or a prediction slice. You can segment your data based on any categorical column or feature in that data (be it for an ML model, data pipeline, data stream, or dataset). For example, you can choose to segment your data by geographies such as “zip code” or “country”, or human attributes like “gender” or “age range”.
Segments are powerful because they allow you to monitor changes in specific categories, classes, or other subsets of your data. They can detect bias or changes that may not be visible when only monitoring the dataset as a whole.
Projects in the Starter and Expert plans have a limited number of Features that can be associated with them. However, our platform can easily scale to monitoring Projects with thousands of features. If you have large ML models or datasets which need monitoring, please contact us to discuss our Enterprise offering.
That's correct! WhyLabs uses the whylogs open source library to collect summary statistics about your data. This approach enables us to provide highly accurate and scalable data and model monitoring, regardless of your data volume.
None at all! WhyLabs relies on statistical profiles of your data generated by our open source library, whylogs. WhyLabs collects none of your data and your data never leaves your environment.
The whylogs library acts as an agent, capturing key statistical properties of data, such as the distribution, the number of missing values, and a wide range of configurable custom metrics. Using these summary statistics, WhyLabs AI Observatory is able to accurately represent the data and help you monitor your data and models.
To learn more about how WhyLabs guarantees your data privacy and security, take a look at our documentation or email us at [email protected].
Absolutely. By using whylogs, you can collect metadata about data regardless of whether it’s being used for an ML model, flowing through a data stream, or being transformed into a data pipeline.
WhyLabs supports a number of different types of monitors that you can configure based on your use case. If you’re monitoring an ML model, you may be interested in tracking data drift or model performance degradation. If you’re monitoring a data pipeline, you may prefer our data quality monitors, such as monitoring for missing values or data type changes.
Regardless of whether you wish to monitor data or models, WhyLabs can be configured to meet your needs.
All kinds!
You can monitor tabular, text, image, embedding, video, or audio data with WhyLabs. If you are interested in other data types, please contact us and we’d be happy to help.
You can also monitor batch or streaming data, regardless of the data volume.
- Classification (single String)
- Regression (single Numeric)
- Embeddings (multiple Numeric)
- Sequence (multiple Strings, multiple Numeric)
- Recommendation system (multiple Strings)
- Bounding box (multiple Numeric, plus maybe one String/Numeric)
- Forecast (multiple Numeric)
Note that any custom output, including multiple output support is available.
Since WhyLabs relies on statistical profiles generated by whylogs, we easily scale to Projects with massive data volume, without having to resort to sampling. whylogs profiles are highly efficient (meaning that they store a lot of information in a small amount of computer memory), so it’s easy to send, store, and analyze them.
Other solutions will require you to sample data in your ML model, data pipeline, data stream, or or dataset. Sampling decreases the effectiveness of monitoring because it loses key statistical information about the dataset. To learn more, check out our Sampling versus Profiling blog post.
Our Starter and Expert customers have two options: opening a support ticket or asking a question in the community Slack. To open a ticket, log in your WhyLabs account. Expand the main menu on the top left-hand side and open “Support Center”. Follow the prompts to submit a support request. To ask a question over Slack, join the WhyLabs Slack Community.
Out of the box, whylogs is available in Python 3, Java, Scala and Apache Spark. In addition, for other languages, WhyLabs supports statistics collection via a REST container.
For the WhyLabs Expert and Enterprise plans we currently accept ACH or wire payments. You can also pay via the AWS Marketplace or reach out to us for a private offer. If you’d like to explore additional payment options, please contact us at [email protected].
Yes, WhyLabs has completed our SOC 2 Type 2 examination with zero exceptions. To request The WhyLabs SOC 2 Type 2 report, please contact your account manager or email [email protected]. To learn more about security at WhyLabs, take a look at https://whylabs.ai/data-privacy.
