Overview
The Evela Health Model is a compact instruction-tuned text model that converts summaries of wearable and phone activity data into structured, per-dimension health readouts. Given a 30-day window of daily step, distance, floors, and optional heart rate signals, it produces an ordinal 1 to 5 level for each health dimension it can assess, with a short rationale naming the signals it used. For dimensions where sensor data carries no signal, the model declines to score and states why, rather than inventing one. The model is approximately 0.5 billion parameters, derived from an Apache-2.0 licensed open-weights base and LoRA fine-tuned on synthetic training pairs sampled from aggregate statistics of a production wearable corpus. No patient records are used in training, and no personal health data is embodied in the weights. At ~1 GB fp16, it serves comfortably from a single instance and is intended for high-volume summarization workloads where routing continuous wearable data through a general-purpose frontier model is prohibitively expensive. Input is a plain-text device summary; output is a plain-text structured readout, both over the standard Bedrock chat interface. Intended use is wellness and activity summarization for health and fitness applications. The model is explicitly not a medical device and must not be used for diagnosis, treatment, or clinical decision making. Scoring is grounded in Evela's eight-dimension preventive-health framework, which the model applies consistently and conservatively: it will refuse to score sleep, nutrition, prevention, sexual health, skin and hair, and social connections from sensor data alone, since these dimensions require additional structured input beyond what wearables and phone sensors capture.
Highlights
- Compact and cost-efficient A 0.5 billion parameter model that serves from a single instance, keeping the cost floor an order of magnitude below routing wearable data through a general-purpose frontier model.
Details
Features and programs
Financing for AWS Marketplace purchases
Pricing
Dimension | Description | Cost |
|---|---|---|
ml.g4dn.xlarge Inference (Batch) Recommended | Model inference on the ml.g4dn.xlarge instance type, batch mode | $1.12/host/hour |
inference.count.m.i.c Inference Pricing | inference.count.m.i.c Inference Pricing | $1.12/request |
Vendor refund policy
All software charges are final and non-refundable. Billing is per instance-hour and the buyer controls when charges stop by deleting the deployed endpoint. For support, contact support@evela.health .
How can we make this page better?
Legal
Vendor terms and conditions
Content disclaimer
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Version release notes
8 health dimensions with activity data sets.
Additional details
Inputs
- Summary
The model accepts a standard chat-formatted JSON payload over the Bedrock Converse or SageMaker runtime API. Each request contains a system message carrying the Evela instruction set and a user message carrying a plain-text summary of a 30-day activity window. The user message follows a fixed text structure:
Period: 30-day window Days with data: of 30 (
% coverage) Data sources: <device, fitbit, oura, comma-separated> Heart rate available: <yes|no>
Daily averages: Steps: steps Distance: meters Floors climbed: floors Steps interquartile range: - [optional]
Heart rate: [only when available] Resting (estimated): bpm Daily average: bpm Peak: bpm
- Input MIME type
- application/json
Support
Vendor support
For support send emails to support@evela.health
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

