Our Technology
1. Vor Engine
Approach
Our Engine represents a groundbreaking approach to the automation of fitness and health routines. Combining several different non-deterministic algorithms into a single streamlined process, the Engine quickly and reliably generates personalised outputs with full transparency and explainability.
Data
The Vor Engine utilises a combination of static and dynamic data, including:
- Data model representation of the body, including muscle groups, individual muscles, tendons, muscle relationships like antagonist/agonist, and muscle size
- A database of exercises and exercise attributes including movement data, form complexity, required equipment, etc.
- Gym information including policies, equipment, layout
- Exerciser information such as experience, goals, injuries, and exercise history
While not in database form, the Engine was designed and built on a foundation of exercise science expertise and best practices. As new research is published, the team at Vor updates the Engine to ensure its processes are informed and backed by science.
Safety and ethics
Unlike Chat GPT and other large language models (LLMs), the Vor Engine offers its users complete explainability and traceability of its outputs. Informed by experts, the Engine also contains built in risk mitigation policies and harm prevention boundaries.
The Vor Engine, by design, requires a human in the loop. This provides accountability and an additional layer of risk mitigation.
Access
The Vor Engine is primarily accessed via a GraphQL API, which is intended for use in client integrations. However, it’s core functionality is not tied to GraphQL, and a REST API or gRPC API can be implemented based on client needs.
API
The Vor Engine provides basic create, read, update, and delete (CRUD) functionality for all of its resources, e.g. equipment, exercises, gyms, etc. The resource graph is fully explorable, enabling a user to query for not only a specific item, but also its contextual connections. The endpoints also provide access to the static data models and resources used by the Engine, such as the muscle data model.
For programme generation in particular, the Engine provides several endpoints. Programme generation in the Engine is done in a work queue, which is split between scheduling and fetching the status and result of a programme generation job. This separation between scheduling and result frees any future front-end or integration from keeping long-running connections, and the API remains fast and responsive no matter the load on the job queue.
Additionally, it provides a validation input endpoint that is designed to be very responsive and safe to call repeatedly from interactive front-ends, which require real time feedback.
Integration Partners
While a front-end component of the Vor Engine is available for demo, it is not intended for use in production or integration. Instead, Vor partners with existing front-end products to offer a packaged solution for clients that do not have an app or front-end of their own. Vor remains open to future collaborations and partnerships for app/front-end integration.
Vor also provides an exercise database, hosted in Google Cloud Platform, but can integrate with a client’s existing exercise database if needed.
.svg)
2. Trainer Interface
Our Engine is the heart of Vor tech, but without a trainer interface it is just lines of code.
Our Engine is designed to seamlessly adapt to your existing interface through API integration, but we also work with trusted partners to offer an intuitive trainer interface that is designed to make life easier for your PTs - allowing them to quickly and easily generate, tweak and deliver their client workouts.
We can also provide an exercise database, hosted in Google Cloud Platform, or we can integrate with your existing exercise database, which is often a good choice if you have a unique movement or exercise methodology.
Overview:
- Option to customise and add your brand.
- Intuitive design allows trainers to interact quickly and easily.
- Or seamlessly integrate Vor into your existing trainer app.
3. Exerciser Interface
The way your members see you is important. So is slick end-to-end usability.
If you want to offer a way for members to access workouts on their own, we partner with existing front-end products to offer a packaged solution for our clients that do not have an app or front-end of their own.
If you have this in place already, we can quickly and easily integrate with your in-house tech, a third party, or we can set you up with one of our front end app partners. It’s up to you.
We can also aggregate member feedback to their workouts, in order to further optimise your version of the Vor engine.
Overview
- You may not need a consumer interface.
- If you do, we can work with your in-house tech, or introduce the right partner.
- Member feedback can be fed back into your Vor engine.
Why use Vor?
Vor’s adaptive AI engine uses principles from astrophysics to solve the challenges of scale, safety and accessibility in the fitness space.
Using a suite of parameters about individual exercisers, such as health metrics, goals and injuries, the engine generates appropriate and highly effective personalised workout plans in seconds.
Vor is built to adapt to your unique training style, equipment, gym floor plan(s) and other critical brand differentiators. No two instances are the same.
You can put the tech in the hands of trainers, or provide workout routines directly to exercisers with varying degrees of trainer supervision.
Unlike Large Language Model (LLMs), which are built to mimic natural human language, Vor’s unique AI engine has governance, oversight and is expertly designed by qualified personal trainers.
There is no erroneous data or guesswork within the Vor engine - every workout generated is reliable and traceable.
See Vor in action
FAQs
We know that personal trainers add huge value to gyms and the fitness industry, and many of their skills can't be replaced by technology.
Vor automates the workout creation portion of a trainer's role. It can be used by personal trainers in planning workout programmes and classes. Alternatively, Vor can be used directly by exercisers with varying degrees of trainer supervision, customisation and control.
No. When it comes to your tech stack, Vor can integrate with your internal or external applications via API.
For example, Vor workouts can be delivered via an existing member app directly to exercisers. Or, Vor can be made available via an existing trainer portal to assist teams with workout programming.
If you don't have a tech stack in place, Vor can be accessed by a simple web application from any device.
From a gym equipment perspective, Vor creates workouts customised to your available equipment and gym floor layout. You don't need to invest in any new gym equipment to embed Vor into your operations.
Vor has been intentionally designed by our experienced team of AI experts and personal trainers every step of the way. Rooted in exercise science best practices, it uses parameters for injuries, targets, training environments and backgrounds to make sure that workouts are safe, appropriate and effective for exercisers.
There is no erroneous data or guesswork within the Vor engine - every workout generated is reliable, explainable and traceable.
Our automation engine is not a plugin to a Large Language Model (LLM) like ChatGPT, which provides unreliable and unpredictable results that might look good enough to the untrained eye, but aren’t safe or appropriate for exercisers.
We have been pioneering our unique technology since 2021 and it is engineered specifically for physical exercise. LLMs, on the other hand, are designed to mimic human language, so aren't trained in exercise science and biology.
Today, our tech benchmarks at 100-200 X faster workout generation for personal trainers when compared with writing personalised workout programmes manually. This is a hugely significant time and cost saving for your bottom-line.
