Built for your developers.
01
Connect
Share runs, feedback, availability and goals for each user who opts in.
02
Request
Ask for a runner’s upcoming slots: today’s session and the ones after it.
03
Render and sync
Show it in your app and send it to the watch as a structured workout.
04
Send results back
Return the FIT file and feedback after each run. The next session updates.
Technical documentation and support come with every package.
How the data moves
Run Specific Model, cheaper and outperforms LLMs
LLMs are not the AI solution to every problem. They are extremely good at language prediction but repeatedly, domain specific models beat LLMs in domain specific tasks for less than 1% of the computational cost.
Training sessions that are optimal and flexible
One for every user, every day, and rewritten when their week changes.
No runner stuck
on a generic plan.
of Running Algorithm users have never had two runs in a row that were too hard. Adaptive plans keep people consistent, and consistent users stay.
01
FIT files
Every run your device records: pace, heart rate, power and terrain.
02
Feedback
How the run felt, in one tap. Our fatigue model learns from every answer.
03
Availability
When they can train this week, and for how long.
04
Goals
The race, the distance or the habit they are working towards.
until most users start feeling faster on their runs
of runs rated “Just right” by the runner
miles run on sessions from our engine
1:01:37
half marathon by Jimmy Whelan, ex-WorldTour pro, who trains with our engine
Send
Pass us the runs your devices already record, plus the user’s feedback, availability and goal.
Receive
Get each session back as structured data: steps, targets and timing, ready to map to your own workout format.
Adapt
Every new run and every change of plan updates what comes next, automatically.
1
Slots
A list of upcoming sessions. When the runner’s week changes, the engine rewrites the slots ahead.
2
Session
A name and coaching notes your app can show the runner as they are.
3
When
The day and start time, fitted to the runner’s availability.
4
Structure
Warm-up, repeated blocks and cool-down, built from components.
5
Effort
How far each component goes and which training zone it sits in.
6
Targets
A pace range for every component, also given as speed in metres per second.
Delivered as structured data your team maps to its own workout format. Technical documentation is shared with launch partners.
BUILT FOR
Built for
health data.
Used only to prescribe sessions for your users. Never sold, never used for advertising.
Runners opt in before anything is shared with the engine, and can withdraw at any time.
Processed under UK GDPR, with data minimised to what the engine needs.
Build with us