Sold by: Salient Engineering
Deployed on AWS
Free Trial
A self-hosted, production-ready Strands Decider E2B model deployed into your AWS environment with a single click. Because everything runs entirely within your private cloud, your data stays secure, isolated, and fully under your control. Best of all, unlimited usage of a dedicated decision model.
Overview
This is a self-hosted deployment of the Strands Decider E2B decision model. It runs as a single EC2 instance so your data stays in your account. Access is via HTTP on port 8000. Once the instance is powered on, the server requires up to 40 minutes on the first boot to load the model before it is ready to serve requests. A later start is faster: about 2 minutes on g5.xlarge, and under a minute on g4dn.xlarge once the volume is warm. The CPU size, c7i.4xlarge, took about 9 minutes on first boot. Highlights of the Strands Decider model include:
-
A decision model, not a text generator. It answers a choice, a yes/no (noul), or an ordered score, and every answer carries a calibrated confidence. One request can ask several questions about the same state.
-
Built on google/gemma-4-E2B-it. This v1 checkpoint is a LoRA adapter plus a small readout head, a soup of three training seeds.
-
Published scores on the model card: JevBench public at a 4,096-token window, 180 of 231 tasks right, Brier 0.313, ECE 0.053. Decision Index 0.2.1 balanced skill at a 32,768-token window is 29.78; the card notes that figure was measured before an E2B attention fix and may be off. The server on this image uses the 4,096-token window.
-
Intended for the decisions inside an agent: model routing, tool selection, argument checking, triage, guardrails, and evaluations. Long, multi-step documents are the weak spot called out on the model card.
Choose the instance type when you launch. g5.xlarge (NVIDIA A10G, 24 GB of GPU memory) is the recommended size for production throughput. g4dn.xlarge (NVIDIA T4, 16 GB of GPU memory) is the lower-cost GPU; the loaded weights use about 9.3 GB. c7i.4xlarge is the CPU-only size. On CPU the weights are fp32, about 19 GB, which is why this image uses c7i.4xlarge rather than a smaller c7i size.
Example API calls (replace PUBLIC_IP with the instance address):
Health check:
curl http://PUBLIC_IP:8000/health
Expected response includes status ok and a device of cuda or cpu.
Decide:
curl -X POST http://PUBLIC_IP:8000/v1/systemone -H 'Content-Type: application/json' -d '{"state":"Help! My payouts have been failing for 3 days!","questions":{"is_urgent":{"type":"noul","instructions":"Does this convey urgency?"}}}'
Highlights
- Data security, privacy, and confidentiality
- Predictable cost
- Unlimited usage of a dedicated model
Details
Sold by
Categories
Delivery method
Delivery option
64-bit (x86) Amazon Machine Image (AMI)
Latest version
Operating system
Ubuntu 26.04
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
Free trial
Try this product free for 5 days according to the free trial terms set by the vendor. Usage-based pricing is in effect for usage beyond the free trial terms. Your free trial gets automatically converted to a paid subscription when the trial ends, but may be canceled any time before that.
Pricing is based on actual usage, with charges varying according to how much you consume. 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.
Dimension | Description | Cost/hour |
|---|---|---|
g5.xlarge Recommended | Production GPU, NVIDIA A10G, 24 GB | $0.09 |
g4dn.xlarge | Lower-cost GPU, NVIDIA T4, 16 GB | $0.09 |
c7i.4xlarge | CPU-only, 16 vCPU, 32 GB memory | $0.09 |
Dimensions summary
You pay by the hour based on the compute instance you run this decision model on. Three options let you match hardware to your workload. The g5.xlarge uses a production GPU with 24 GB. The g4dn.xlarge uses a GPU with 16 GB. The c7i.4xlarge is CPU-only, with 16 vCPUs and 32 GB of memory. These are independent choices, not tiers of service. All run the same model. Your cost scales with how many hours you keep an instance running, so you can stop instances to control spend.
Top-of-mind questions for buyers
Each option is a separate instance type you pay for by the hour. The g5.xlarge gives you a production GPU with 24 GB. The g4dn.xlarge gives you a GPU with 16 GB. The c7i.4xlarge is CPU-only, with 16 vCPUs and 32 GB of memory.
Hourly charges apply only while an instance runs. Stopping it halts software charges, though AWS storage fees may still apply. The GPU options suit faster inference workloads. The CPU-only option runs the same model without a GPU. You pick one instance per running workload.
The hourly charge is for running the chosen compute instance. The model is a decision model that picks between options and rates things with a calibrated confidence score. It is released under an Apache-2.0 license. All three instance options run the same model.
huggingface.co
Helpful?
Vendor refund policy
Refunds may be considered on a per-case basis. Please contact us at support@salientengineering.com for inquiries.
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
64-bit (x86) Amazon Machine Image (AMI)
Amazon Machine Image (AMI)
An AMI is a virtual image that provides the information required to launch an instance. Amazon EC2 (Elastic Compute Cloud) instances are virtual servers on which you can run your applications and workloads, offering varying combinations of CPU, memory, storage, and networking resources. You can launch as many instances from as many different AMIs as you need.
Version release notes
Configured for production environments. On the first boot, please allow up to 40 minutes for the model to load. A later start is faster: about 2 minutes on g5.xlarge, under a minute on g4dn.xlarge once the volume is warm, and about 9 minutes on the first boot of c7i.4xlarge. The decision API is exposed on port 8000.
Health: curl http://<PUBLIC_IP>:8000/health
Test via HTTP with: curl -X POST http://<PUBLIC_IP>:8000/v1/systemone -H 'Content-Type: application/json' -d '{"state":"Help! My payouts have been failing for 3 days!","questions":{"is_urgent":{"type":"noul","instructions":"Does this convey urgency?"}}}'
Additional details
Usage instructions
- Deploy the EC2 instance. g5.xlarge is the recommended production size (NVIDIA A10G). g4dn.xlarge is the lower-cost GPU (NVIDIA T4). c7i.4xlarge is the CPU-only size. Configure the Security Group to only allow inbound ports 22 and 8000 from your trusted IP address(es).
- After the instance is powered on, allow up to 40 minutes on the first boot for the server to load the model. GET /health returns 200 when ready. Later starts are about 2 minutes on g5.xlarge and under a minute on a warm g4dn.xlarge. c7i.4xlarge took about 9 minutes on first boot.
- Access Strands Decider E2B via HTTP on port 8000. POST /v1/systemone accepts a state and typed questions (noul, choice, score). GET /health reports readiness and the device (cuda or cpu).
Test via HTTP with: curl -X POST http://<PUBLIC_IP>:8000/v1/systemone -H 'Content-Type: application/json' -d '{"state":"Help! My payouts have been failing for 3 days!","questions":{"is_urgent":{"type":"noul","instructions":"Does this convey urgency?"}}}'
Resources
Vendor resources
Support
Vendor support
The Salient Engineering support team can be reached at: support@salientengineering.com
Our team is happy to assist with deployment and configuration issues.
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
Send a ticket, log line, document or JSON object plus typed questions and get back a choice, a yes/no probability or a score level, never generated text. Runs inside your AWS account on a SageMaker endpoint, so no text is sent to a third-party API.
The fast, low-cost Zuse: a European decision model that works like TypeSafe's Jev and runs in your own AWS account. Give it a document and your questions; it picks an answer from your options and says how sure it is. 83 languages, 27 European ones trained. Your documents never leave your VPC.
Opt-in consumer data for Australia. Free data sample available, please contact us to get full access to the data sets.
Opt-in consumer data, available for all major regions and nationwide across New Zealand. Free data sample available, please contact us to get full access to the data sets.
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.