• English
  • Русский(soon)
  • 한국어(soon)

About Actiq

Build Healthy Habits with AI

Actiq is gamified training app powered by AI sport coaches for athletes. Actiq disrupt sport coaching by using the fusion of large language models and computer vision. By adding AI twin coaches into training practice, human coaches can radically transform their training process and coach athletes autonomously and remotely. Athletes can obtain monetary rewards by completing training blocks with AI. Supported sports: tennis, golf, power training, running, cycling, swimming, yoga, rehabilitation gymnastics and many others.

AI Coach is an LLM agent that can literally communicate with athlete to establish constant Human Feedback Learning to customize training plan in real-time. To build or update initial training plan, AI coach uses the outputs of deep learning neural network based on Strava and other health-related data sources plus to Computer Vision data. Training plan includes blocks such as indoor gamified workouts or outdoor quests. AI coach will rate athletes and reward them with tokenized incentives accordingly.

Actiquest game provides you chances to work out on a daily basis either in single or world challenge mode. It also record your training results and convert athletes perfomance into valuable rewards.

Actiquest generally aims to incentivize fitness and sport, helping to reduce obesity and sedentariness, and giving users an opportunity to earn an income for improving their health.

Actiquest will bridge data from all professional sports applications into one place and give you rewards for tracking user activities. Strava, Zwift, Peloton, RGT and other platforms users can connect their profiles to the Actiq App and get tokenized incentives on AI
fair demand.

Key points

Roadmap


R&D Deep Learning Algorithms

• GAP (Grade Adjusted Pace) DL model + test
• Data Sourcing (Health Datasets)
• Recovery model (DL) + test
• Strava API <=> Langchain <=> LLM API

Deep Learning Training Plan

R&D LLM Works

•  VoxelGPT + Training
•  Langchain + Decision Intelligence + training
•  VoxelGPT + Whisper
•  VoxelGPT + Reasoning model
•  VoxelGPT + various NLP Tests

LLM AI Coaches

AI DAO

•  $ACTI Treasury Smart Contract
•  AI<=>DAO Voting Oracle
•  AI<=>Web3 Incentives Oracle Concept
•  AI DAO Concept

DAO Oracles

Other Developments

•  Actiq Web3 Dashboard UI/UX prototyping
•  Gamification Concept (indoor/outdoor sports)
•  Marketing activities
•  Content production

Web3 Alliances

MVP Development (phase 0) Q3 Q4

•  Custdev campaign (200 Triathletes)
•  Personalized Training Plan (Omnisport Model)
•  Enhanced dataset (Sources: Strava, Garmin, Google Medpalm, Pose Estimation)
•  Actiq Early Bird Program (AI Training for PreTokens)
•  AI by AI Concept

ACTIQ

Integrations Q3 Q4

•  Garmin Health SDK
•  Strava API v3

API Vendor Integrations

ACTIQ 2nd Stage Developments

•  Push notifications on user activities
•  Phygital Assets and Collectibles Concept
•  Challenge Mechanics
•  Web3 Leaderboard Concept

ACTIQ

Sport SSI pilot

•  A decentralized identity that can be integrated into fitness apps.

Blockchain SSI

Langchain agents

•  LC-Pose
•  LC-NLP
•  LC-Training Plan

Integrations

MVP Development (phase 1)

•  Epsilla VectorDB performance tests
•  App UI/UX
•  App User Flow
•  Training Plan Dataset Update
•  LLM’s tests
•  СV Pose Estimation model tests

ACTIQ

MVP Development (phase 2) Q1 Q2

•  MVP architecture ready
•  Epsilla 2 SQL mod
•  Yolo v8 Pose NAS mod (1:1 person, mobile version)
•  VoxelGPT mod (fastener)
•  3D measuring algorithm with Yolo-dist library
•  Fifty-One-auto-labeling-mod (Yolo2LLM seamless integration)
•  Sports type recognition (Yolo v.8) + Sports Objects recognition

ACTIQ

Integrations Q1 Q2

•  Tennis Model Labeling
•  Golf Model Labeling
•  Padel Model Labeling
•  VoxelGptSport training

Labeling AI training

AI Sport Coach Developments

•  Coach visualization wizard (Sport Coach avatar realtime design/producing using picture/propmt/Controlnet/Stable Diffusion)
•  Yolo Pose mod (FPS-agnostic player tracking model - golf & tennis)
•  100 athletes custdev Campaign (Tennis) - AI training
•  100 athletes custdev Campaign (Golf) - AI training

AI Coach

Self-Learning & Auto Labeling

•  Testing Auto-Labeling CV outputs for play sports.

Voxel51 CV2LLM

Actiquest brand guide

Here you can find all the information you need about logos, colors, gradients, and even the structure of SMM posts

Coming Soon

ACTIQ
Litepaper

ACTIQ Litepaper & General Information for investors and partners

v1.0 Updated 20.05.2023

ACTIQ Blog

Actiquest "AI Jorney" Blog
v1.0, updated weekly