Effortlessly deploy machine learning models and applications
If you’ve reached the end of an AI project cycle, putting your machine learning models or applications into production, you know that this resource-intensive stage of industrialisation can be a real challenge. To facilitate this, deploy your Docker images effortlessly and without Kubernetes expertise using AI Deploy. Carry out requests via API for your models and via the web interface for your production applications, while we manage the infrastructure and security of the environments.
Speedy managed deployment
Switching from a machine learning prototype to deploying a model into production is often a time-consuming process. Use AI Deploy from your Control Panel, via the API or in the command line (CLI), and easily industrialise your models with flexibility in a matter of minutes.
No architecture to manage
Export your models or applications in a Docker image, and AI Deploy will take care of the rest. Your deployments are supported with total abstraction of the hardware architecture.
Flexibility and performance
Specify a minimum and maximum number of instances for your deployments, and only pay when you use them. AI Deploy uses automatic scaling. Whether you have 10 requests per day or 10,000 per minute, we will increase and decrease the resources you need to give you an optimal experience.
Our sovereign European cloud ensures that your data is secure. Our cloud infrastructures and services are ISO/IEC 27001, 27017, 27018 and 27701 certified. With our health data hosting compliance, you can host healthcare data securely.
CPU and GPU resources
Deploy models and applications with NVIDIA CPUs or GPUs according to your needs, even for the most demanding inferences.
Select deployments on multiple instances to benefit from high availability. Load balancing is automatic and managed by AI Deploy.
Use cases for AI Deploy
Startups and SMEs
Are you working on internal projects that you want to deploy to production? It has never been so simple and achievable for your teams. Streamlit, Gradio, or simply API access points in a Docker image: all your projects can be put into production smoothly and easily. Your imagination is the only limit!
Automate your deployments with OVHcloud APIs. Securely provide your customers with individual AI models and applications, and keep predictive control of access and costs.
Business dashboard, fraud analysis, and more. Whatever your use case, easily deploy your AI projects to production for you or your customers, without using a team of architects. Control your budget with predictive pricing.
High availability and available resources
By deploying multiple instances, you can ensure high availability for your infrastructure. You can also choose the right computing power needed.
Thanks to our rolling upgrade mechanism, deployment upgrades are performed without any downtime. So you can keep working on your modelling, and keep production versions up-to-date.
Whether your model receives a high volume of requests or you use it at specific times of day, we automatically scale its deployment so it adapts in record time.
Metrics and logs
With quick access to your event logs, you can easily monitor your tasks. You can track the number of calls or even the latency time.
Protect your deployment environment
Using an application token for access, you can be sure that only your authorised employees can access AI Deploy. Select and set these tokens to apply user privileges as needed.
Simple, pay-as-you-go rates
You can opt for pay-per-use billing with AI Deploy. For each deployment you launch, you only pay for the resources of the instances you use (GPU or CPU). The number of calls is unlimited, creating predictable costs with no commitment required.
Managed from the OVHcloud Control Panel, via the API or in the command line
Depending on your skills and preferences, you can launch and track your deployments from the web interface, via the API, or via the command line, no matter what programming language you use.
Train your AI, machine learning and deep learning models efficiently and easily, and optimise your GPU usage.