Sold by: DEEPX CO.,LTD.
Deployed on AWS
DEEPX Greengrass Solution is a one-click CloudFormation product that compiles ONNX models to DEEPX NPU binaries and deploys the full NPU runtime (driver, firmware, dx_rt, dx_stream) to AWS IoT Greengrass edge fleets. It bridges model compilation and edge inference for DEEPX M1 NPUs, automating runtime install and firmware flashing across your core devices. Keywords: DEEPX NPU, AWS IoT Greengrass, edge AI inference, ONNX compiler, NPU runtime, dx_rt.
Overview
DEEPX Greengrass Solution delivers an end-to-end path from trained model to edge inference through a single AWS CloudFormation deployment. It launches the DEEPX Compiler pipeline, spinning up the DEEPX Compiler AMI to convert your ONNX models into optimized .dxnn binaries for DEEPX NPUs, and, in the same template, provisions the AWS IoT Greengrass infrastructure (thing group, token-exchange role, and a managed deployment) that rolls the DEEPX NPU runtime out to your edge fleet. Model compilation and edge runtime provisioning are unified in one deliverable, so you go from ONNX to inference-ready devices without stitching together separate tools. On the compile side, the pipeline runs dxcom on the DEEPX Compiler AMI and writes the resulting .dxnn artifacts back to your S3 bucket. On the edge side, the com.deepx.dx-runtime Greengrass component installs and verifies the full runtime stack on every core device in the correct order: the NPU Linux driver (DKMS), device firmware, the dx_rt runtime, and the dx_stream GStreamer pipeline. Installation is idempotent: the component detects the currently installed driver and firmware versions and skips any target already up to date, so firmware flashing, driver builds, and runtime compilation run over-the-air, safely, across redeployments and fleet-wide version upgrades on both aarch64 (e.g. Raspberry Pi) and x86_64 hosts. The solution ships with a guided web console to compile models, configure the target IoT thing group, select the runtime version, and validate the full deployment before launch. It is built for teams operating DEEPX M1 NPU devices at the edge who need a repeatable, versioned, fleet-scale path from model compilation to deployed inference. Keywords: DEEPX NPU, ONNX compiler, dxnn, AWS IoT Greengrass, edge AI, NPU runtime, over-the-air deployment, firmware flashing, dx_rt, dx_stream, M1 NPU, edge inference.
Highlights
- End-to-end model-to-edge in one deployment: a single AWS CloudFormation template compiles your ONNX models to optimized .dxnn binaries on the DEEPX Compiler AMI and provisions the AWS IoT Greengrass infrastructure that rolls the DEEPX NPU runtime out to your entire edge fleet, with no manual per-device setup.
- Complete NPU runtime, installed and verified over-the-air: the com.deepx.dx-runtime Greengrass component installs the NPU driver, device firmware, dx_rt, and dx_stream in the correct order on every core device, with idempotent version checks that safely skip targets already up to date across redeployments and fleet-wide upgrades.
- Built for DEEPX M1 NPU fleets at scale: guided web console to compile models, target an IoT thing group, pin the runtime version, and validate before launch. Repeatable, versioned edge AI provisioning on both aarch64 (Raspberry Pi) and x86_64 hosts.
Details
Sold by
Categories
Delivery method
Delivery option
DEEPX Compiler AMI + Greengrass (CloudFormation)
Latest version
Operating system
Ubuntu 22.04
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
This product is available free of charge. Free subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Vendor refund policy
This product is offered free of charge. As there are no charges for its use, no refunds are applicable.
For any questions, please contact: info@deepx.ai
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
DEEPX Compiler AMI + Greengrass (CloudFormation)
This Single-AMI CloudFormation delivery packages the DEEPX model-compilation pipeline and the DEEPX edge-runtime rollout into one template, so a single stack deployment takes you from an ONNX model to inference-ready DEEPX NPU devices.
The stack creates a versioned, encrypted S3 bucket (ModelBucketName) for your models. Whenever you upload an ONNX model together with its JSON compile config to the same folder, a Step Functions workflow automatically launches a temporary EC2 instance from the DEEPX Compiler AMI (ImageId) in the VPC and subnet you specify (VpcId, SubnetId, InstanceType), runs dxcom to compile the model into an optimized .dxnn binary, writes the result back to the same S3 folder, and terminates the instance. Compiler instances exist only for the duration of each compile job, so you pay for EC2 only while compiling.
In the same deployment, the template stands up the AWS IoT Greengrass infrastructure: an IoT thing group (ThingGroupName, or a stack-unique -cores when left empty), a least-privilege token-exchange role for the core devices, and a managed Greengrass deployment. If the named thing group already exists, the stack adopts it instead of failing, and rolls the runtime out to the devices already in it.
The Greengrass deployment publishes and rolls out the com.deepx.dx-runtime component (DxRuntimeComponentVersion). On every core device in the thing group, the component downloads its install artifacts from a public S3 base URL (DxRuntimeArtifactBaseUrl) and installs the full runtime stack over-the-air, in order: NPU Linux driver (DKMS), device firmware, dx_rt, and dx_stream. Installation is idempotent: the component detects already-installed targets and skips anything up to date, making redeployments and fleet-wide version upgrades safe and fast.
Prerequisites: a globally unique S3 bucket name that is not yet in use (the stack creates the bucket), a VPC subnet with outbound internet access or VPC endpoints for S3, SSM, and CloudWatch Logs, and one or more registered Greengrass core devices joined to the target thing group with outbound HTTPS access for artifact downloads. Core devices are supported on both aarch64 (e.g. Raspberry Pi) and x86_64 hosts running DEEPX M1 NPUs. All stack behavior is parameter-driven, with no manual SSH or per-device configuration required.
CloudFormation Template (CFT)
AWS CloudFormation templates are JSON or YAML-formatted text files that simplify provisioning and management on AWS. The templates describe the service or application architecture you want to deploy, and AWS CloudFormation uses those templates to provision and configure the required services (such as Amazon EC2 instances or Amazon RDS DB instances). The deployed application and associated resources are called a "stack."
Version release notes
Initial AWS Marketplace release of DEEPX Greengrass Solution. A single CloudFormation deployment compiles ONNX models to .dxnn on the DEEPX Compiler AMI (dxcom) and provisions AWS IoT Greengrass to install the DEEPX NPU runtime across your edge fleet over-the-air.
Bundled DEEPX NPU runtime (com.deepx.dx-runtime v2.4.0):
- NPU Linux driver (DKMS) v2.5.1
- Device firmware v2.7.3 (M1 / mdot2)
- dx_rt v3.4.0 debian package installation
- dx_stream GStreamer pipeline v3.1.0
Additional details
Usage instructions
Usage Instructions - DEEPX Greengrass Solution
This product delivers two capabilities from one CloudFormation stack: (1) Compile - turn ONNX models into DEEPX .dxnn files in your account using the DEEPX Compiler AMI (dxcom); (2) Edge deploy - provision AWS IoT Greengrass V2 so the DEEPX NPU runtime (com.deepx.dx-runtime) is installed on your edge core devices.
Prerequisites: an active subscription to this product; a VPC and subnet with outbound internet access (or VPC endpoints for S3, SSM, EC2, Step Functions, CloudWatch Logs); a globally-unique S3 bucket name that does not exist yet (the stack creates and owns the bucket); for edge deploy, one or more AWS IoT Greengrass V2 core devices with DEEPX M1 NPU hardware, outbound HTTPS access, and passwordless sudo for ggc_user (the runtime installs a kernel driver and flashes firmware). DEEPX NPU hardware is sold separately and is not included in this listing.
How to use (self-service):
- Subscribe on AWS Marketplace and launch the CloudFormation template. Choose your region and provide VpcId, SubnetId, and ModelBucketName; ImageId resolves automatically from your subscription. Optionally set ThingGroupName to an existing IoT thing group - the stack adopts it instead of failing - or leave it empty to create a stack-unique group.
- Compile: upload a model.onnx and its model.json config to the same prefix in the stack's S3 bucket, one model pair per prefix (e.g. models/resnet/resnet.onnx + models/resnet/resnet.json). A temporary compiler instance launches automatically, and the .dxnn appears in the same prefix when compilation finishes. Track progress in Step Functions / CloudWatch Logs.
- Edge deploy: install Greengrass Core on your DEEPX NPU devices using the stack outputs - join them to the thing group (ThingGroupName output) and pass the token exchange role (TokenExchangeRoleName output) to the Greengrass installer. Devices in the group receive com.deepx.dx-runtime automatically and install the NPU driver, firmware, dx_rt, and dx_stream over-the-air. Installation is idempotent, so redeployments skip up-to-date devices.
- Deliver the compiled .dxnn to the edge device and run it on the NPU.
No step requires seller approval; deployment and usage are fully self-service. Full account-level disclosure (resources created, IAM scope, external dependencies) is provided in the product documentation.
Support
Vendor support
Email: tech_support@deepx.ai Support URL: https://deepx.ai/contact-us/ Support description: We are committed to helping our customers succeed with their AI projects. Our support includes:
- Dedicated Technical Assistance: Get expert guidance on a wide range of topics, including product integration, developer questions, and model optimization.
- Comprehensive Resources: We provide access to a developer portal, quick-start guides, and online video courses to help you get up and running quickly.
- Scalable Support Plans: Whether you're a developer prototyping with the DX TechBridge Kit or a business transitioning to mass production, we offer tailored support plans to meet your specific needs and ensure seamless deployment.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
The DEEPX Solution is a software product that offers a streamlined, end-to-end workflow for deploying deep learning models to edge devices. It uses a serverless compilation pipeline to automatically trigger the deployment of NPU packages, enabling businesses to quickly and efficiently manage and remotely trigger AI/ML demos on IoT devices.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.