Prevision.io News: First Ever Pay-As-You-Go AI Management Platform Prevision.io Launches on Google Cloud.

Find here all answers to your questions regarding Prevision.io’s AI management platform! 

AI management platform prevision

There are two opportunities to use Prevision.io.   

  • Visit cloud.prevision.io for the Free Trial (14 days)
  • Link to GCP Marketplace for access to the full version on the Google Cloud Marketplace (GCP). A limited number of free credits are made available upon registration for you to conduct a trial of the offering.  Once the credits are exhausted, a Pay-As-You-Go (PAYG) model kicks in.   

Prevision.io offers a Pay-As-You-Go licensing model. Pricing can be found here

  • Prevision.io offers the ability to create models using our AutoML offering (including the ability to access Notebooks, and an SDK).
  • If your model has been built elsewhere…Great!   Prevision.io supports the ability to host, deploy and monitor 3rd party models.  You can see how to do that on the external model import documentation

In-app support is provided by Prevision.io. You can click the Help button to communicate directly with Prevision.io experts.

The free trial is not meant for production-type work. If you want to test, we recommend using the GCP Marketplace. We are happy to work with you to make your experience as efficient as possible. Let us know how we can help!

No, build your own model or build with Prevision.io…or, both! We are flexible to your needs as they evolve from use case to use case.

No. This is the beauty of Prevision.io. As former data scientists, we’ve tried really hard to take care of the hard stuff that is not your job.

Prevision.io provides a set of API for all its features and can be used with Python or R SDKs in your usual tools.

IT monitoring looks for technical metrics, such as model health, usage, response time and consumption of resources (CPU, RAM, GPU). While Prevision.io offers these capabilities, we complement them with data science monitoring in order to help you understand the model’s behavior. To do so, we provide you with information about concept drift, data drift, data integrity and custom alerting.

With Prevision.io deployments you can update the models in production without downtime. In addition, you can use the Champion/Challenger feature to test a model candidate before you put it in production.

No, Prevision.io connects to various data sources outside of GCP scope (e.g. S3, SQL, ODBC, JDBC, FTP)

  • This is the beauty of Prevision.io.  You only pay for when you are using the offering.  Come and go as you please.
  • If you cancel your GCP registration, your work and data will be deleted.

No issues at all! We have it on our roadmap to get over to the other cloud vendors soon.

In the works…but for now, we are only for use on GCP.

Data is stored in Google cloud storage buckets and production data for monitoring is stored in Google Cloud SQL.

Prevision.io offers a pleasant and frictionless UI that allows clickers to quickly work with the platform. You can import data, experiments, deploy and monitor models without writing a single line of code.

Prevision.io offers Python and R SDKs in order to interact with the platform. They are available through your instance notebook or can be retrieved from our github if you prefer working into your own development environnement (SDK documentation). For the one that uses another language, you can directly access our APIs (API documentation)

Sure, the user documentation is available here.

As of today, Prevision.io offers a default configuration for notebooks. If this one isn’t sufficient for you, feel free to reach us and we will find a solution.

Prevision.io is built around team working altogether. You can manage a team directly within your projects and share the knowledge according to your needs.

Updates are constantly made to the platform (e.g. bug fixes and security fixes). Minor updates occur every 6 weeks with major releases occurring quarterly.

Changelogs are publicly available in the user documentation.

Thanks for the feedback, we would love to hear it and improve our product based on the needs of our users. As of today, contacting us directly within the help button is the best way to contact our experts.