Subnet Directory

With the amount of new subnets being added it can be hard to get up to date information across all subnets, so data may be slightly out of date from time to time

Subnet 58

Dippy Speech

Emissions
Value
Recycled
Value
Recycled (24h)
Value
Registration Cost
Value
Active Validators
Value
Active Miners
Value
Active Dual Miners/Validators
Value

ABOUT

What exactly does it do?

Dippy is one of the world’s leading AI companion apps, with over 1 million users. The app has achieved significant success, ranking #3 on the App Store in countries like Germany and receiving coverage from major publications such as Wired magazine. On average, Dippy users spend over an hour engaging with the app.

Beyond the app, they are also the team behind Bittensor’s Subnet 11, dedicated to building the world’s best open-source roleplay LLM. The open-source miner models developed on Subnet 11 directly power the Dippy app, and they also plan to integrate models from their speech subnet into Dippy to enhance its capabilities further.

Their Dippy Empathetic Speech Subnet is focused on developing the most advanced open-source speech model, designed to enable immersive and lifelike AI interactions. By harnessing the collaborative power of the open-source community, they are addressing the growing demand for AI companions that prioritise natural, speech-first engagement. Their goal is to create a model that delivers deeply personalised, empathetic speech interactions—surpassing the limitations of traditional assistants and proprietary models.

Unlike existing models that rely on reference speech recordings, which restrict creative flexibility, they take an innovative approach by using natural language prompting to define speaker identity and style. This method enables more dynamic and personalised roleplay experiences, fostering richer and more engaging interactions for users.

 

Dippy is one of the world’s leading AI companion apps, with over 1 million users. The app has achieved significant success, ranking #3 on the App Store in countries like Germany and receiving coverage from major publications such as Wired magazine. On average, Dippy users spend over an hour engaging with the app.

Beyond the app, they are also the team behind Bittensor’s Subnet 11, dedicated to building the world’s best open-source roleplay LLM. The open-source miner models developed on Subnet 11 directly power the Dippy app, and they also plan to integrate models from their speech subnet into Dippy to enhance its capabilities further.

Their Dippy Empathetic Speech Subnet is focused on developing the most advanced open-source speech model, designed to enable immersive and lifelike AI interactions. By harnessing the collaborative power of the open-source community, they are addressing the growing demand for AI companions that prioritise natural, speech-first engagement. Their goal is to create a model that delivers deeply personalised, empathetic speech interactions—surpassing the limitations of traditional assistants and proprietary models.

Unlike existing models that rely on reference speech recordings, which restrict creative flexibility, they take an innovative approach by using natural language prompting to define speaker identity and style. This method enables more dynamic and personalised roleplay experiences, fostering richer and more engaging interactions for users.

 

PURPOSE

What exactly is the 'product/build'?

Dippy has already demonstrated strong market traction, securing the #3 spot on the German App Store and surpassing 1 million active users. The platform’s high engagement levels, with users spending over an hour on average per session, showcase its real-world adoption and sustained demand.

 

$TAO Integration

Dippy is not just an app—it is deeply integrated within the Bittensor ecosystem, leveraging open-source, decentralised intelligence for a more advanced AI experience:

  • Subnet 11 powers the backend, providing the foundational LLM for AI interactions.
  • Subnet 58 is dedicated to developing cutting-edge AI speech models for enhanced voice interactions.
  • An open-source future, ensuring transparency, innovation, and community-driven development.
  • Network-enhanced capabilities, leveraging Bittensor’s decentralised infrastructure for scalability and efficiency.

 

Breaking Down the Edge

Proven Demand

The app’s organic growth is evidence of its strong market fit:

  • 500,000 monthly active users, engaging consistently with AI companions.
  • 50,000 daily active users, showcasing a high level of retention and habitual use.
  • 1-hour average session length, demonstrating deep user engagement.
  • Sustained natural growth, achieved without reliance on aggressive marketing tactics.

 

Bittensor-Powered Solution

Dippy’s integration within Bittensor provides key advantages over traditional AI models:

  • Decentralised model, reducing reliance on closed, centralised systems.
  • Community-powered development, driving continuous improvement through open collaboration.
  • Open-source infrastructure, ensuring transparency and accessibility.
  • No restrictions, allowing for innovation and adaptability without corporate gatekeeping.

 

Think Deeper: The Power Play

Integration Plan

Dippy is strategically designed to leverage multiple Bittensor subnets for a synergistic AI experience:

  • Subnet 11 as the foundational base model for AI interactions.
  • Subnet 58 as the next-generation voice model for expressive, human-like speech.
  • Cross-subnet synergy, allowing different AI models to interact and enhance each other.
  • Network effects, where the adoption of one feature amplifies the value of the entire system.

Real Adoption

Unlike theoretical AI projects, Dippy has already achieved real-world success:

  • A top-ranking app, already in the hands of millions.
  • Mass market-ready, with proven demand and scalable infrastructure.
  • Validated by real users, not just early adopters or niche communities.
  • Market-proven, demonstrating clear traction and growth potential.

 

The Most Important Part:

Dippy is not just another AI app—it represents the mass adoption of AI, bringing decentralised intelligence to a mainstream audience.

1. Market Validation

  • Top App Store rankings, proving its competitive edge.
  • A real, engaged user base, unlike speculative projects.
  • High levels of active engagement, showing sustained interest.
  • A proven model, that works in real-world applications.

2. Growth Drivers

Dippy’s continued success is powered by:

  • Subnet integration, allowing AI models to evolve and improve.
  • Voice capabilities, unlocking more immersive, human-like interactions.
  • Network effects, where growth in one area benefits the entire ecosystem.
  • Mass appeal, making AI companions more accessible to a global audience.

 

Key Takeaways

  1. Dippy has already succeeded in the real world, with a massive, engaged user base.
  2. It is a market leader, proving demand for AI companions.
  3. Its deep integration with $TAO strengthens its future potential in decentralised AI.
  4. It is future-ready, built for scalability and ongoing innovation.

 

Overview of Miner and Validator Functionality

Miners will leverage existing frameworks to fine-tune models, enhancing the current state-of-the-art open-source TTS model. The refined model weights will then be submitted to a shared Hugging Face repository.

Validators will evaluate and rank model performance using a structured protocol, assessing factors such as naturalness, emotion matching, and clarity. They will have access to a suite of advanced testing and benchmarking tools, utilising state-of-the-art datasets to ensure rigorous assessment.

 

Dippy has already demonstrated strong market traction, securing the #3 spot on the German App Store and surpassing 1 million active users. The platform’s high engagement levels, with users spending over an hour on average per session, showcase its real-world adoption and sustained demand.

 

$TAO Integration

Dippy is not just an app—it is deeply integrated within the Bittensor ecosystem, leveraging open-source, decentralised intelligence for a more advanced AI experience:

  • Subnet 11 powers the backend, providing the foundational LLM for AI interactions.
  • Subnet 58 is dedicated to developing cutting-edge AI speech models for enhanced voice interactions.
  • An open-source future, ensuring transparency, innovation, and community-driven development.
  • Network-enhanced capabilities, leveraging Bittensor’s decentralised infrastructure for scalability and efficiency.

 

Breaking Down the Edge

Proven Demand

The app’s organic growth is evidence of its strong market fit:

  • 500,000 monthly active users, engaging consistently with AI companions.
  • 50,000 daily active users, showcasing a high level of retention and habitual use.
  • 1-hour average session length, demonstrating deep user engagement.
  • Sustained natural growth, achieved without reliance on aggressive marketing tactics.

 

Bittensor-Powered Solution

Dippy’s integration within Bittensor provides key advantages over traditional AI models:

  • Decentralised model, reducing reliance on closed, centralised systems.
  • Community-powered development, driving continuous improvement through open collaboration.
  • Open-source infrastructure, ensuring transparency and accessibility.
  • No restrictions, allowing for innovation and adaptability without corporate gatekeeping.

 

Think Deeper: The Power Play

Integration Plan

Dippy is strategically designed to leverage multiple Bittensor subnets for a synergistic AI experience:

  • Subnet 11 as the foundational base model for AI interactions.
  • Subnet 58 as the next-generation voice model for expressive, human-like speech.
  • Cross-subnet synergy, allowing different AI models to interact and enhance each other.
  • Network effects, where the adoption of one feature amplifies the value of the entire system.

Real Adoption

Unlike theoretical AI projects, Dippy has already achieved real-world success:

  • A top-ranking app, already in the hands of millions.
  • Mass market-ready, with proven demand and scalable infrastructure.
  • Validated by real users, not just early adopters or niche communities.
  • Market-proven, demonstrating clear traction and growth potential.

 

The Most Important Part:

Dippy is not just another AI app—it represents the mass adoption of AI, bringing decentralised intelligence to a mainstream audience.

1. Market Validation

  • Top App Store rankings, proving its competitive edge.
  • A real, engaged user base, unlike speculative projects.
  • High levels of active engagement, showing sustained interest.
  • A proven model, that works in real-world applications.

2. Growth Drivers

Dippy’s continued success is powered by:

  • Subnet integration, allowing AI models to evolve and improve.
  • Voice capabilities, unlocking more immersive, human-like interactions.
  • Network effects, where growth in one area benefits the entire ecosystem.
  • Mass appeal, making AI companions more accessible to a global audience.

 

Key Takeaways

  1. Dippy has already succeeded in the real world, with a massive, engaged user base.
  2. It is a market leader, proving demand for AI companions.
  3. Its deep integration with $TAO strengthens its future potential in decentralised AI.
  4. It is future-ready, built for scalability and ongoing innovation.

 

Overview of Miner and Validator Functionality

Miners will leverage existing frameworks to fine-tune models, enhancing the current state-of-the-art open-source TTS model. The refined model weights will then be submitted to a shared Hugging Face repository.

Validators will evaluate and rank model performance using a structured protocol, assessing factors such as naturalness, emotion matching, and clarity. They will have access to a suite of advanced testing and benchmarking tools, utilising state-of-the-art datasets to ensure rigorous assessment.

 

WHO

Team Info

The team from Impel includes members with backgrounds from prestigious companies like Microsoft, IBM, and Twitter, who have collectively created apps with over 100 million downloads that were groundbreaking in the AI space.

The Impel team is focused on creating AI models that not only offer utility but also incorporate emotional intelligence elements like compassion, empathy, and humor to enhance user engagement.

After experiencing success with the viral app “Wombo,” the Impel founders decided to shift their focus to creating AI products with a proactive and context-aware approach, aiming to cater to billions of consumers.

Despite being a young company founded in August 2023, impel quickly secured a significant 2.1 million preseed funding round within just a month of incorporation, showcasing rapid growth and potential.

Akshat Jagga – CEO

Angad Arneja – COO

The team from Impel includes members with backgrounds from prestigious companies like Microsoft, IBM, and Twitter, who have collectively created apps with over 100 million downloads that were groundbreaking in the AI space.

The Impel team is focused on creating AI models that not only offer utility but also incorporate emotional intelligence elements like compassion, empathy, and humor to enhance user engagement.

After experiencing success with the viral app “Wombo,” the Impel founders decided to shift their focus to creating AI products with a proactive and context-aware approach, aiming to cater to billions of consumers.

Despite being a young company founded in August 2023, impel quickly secured a significant 2.1 million preseed funding round within just a month of incorporation, showcasing rapid growth and potential.

Akshat Jagga – CEO

Angad Arneja – COO

FUTURE

Roadmap

They recognise the complexity of developing a state-of-the-art speech model, so they are structuring the process into three distinct phases.

Phase 1:

  • They are launching a subnet with a robust pipeline for roleplay-specific TTS models, designed to interpret prompts for speaker identity and stylistic speech descriptions.
  • They are introducing an infinitely scalable synthetic speech data pipeline.
  • They are implementing a public model leaderboard, ranking submissions based on core evaluation metrics.
  • They are incorporating Human Likeness Score and Word Error Rate as live evaluation criteria for continuous model assessment.

Phase 2:

  • They will refine their TTS models to produce more creatively expressive and highly human-like speech outputs.
  • They will showcase the highest-scoring models, making them publicly accessible through their front-end interface.

Phase 3:

  • They aim to advance toward an end-to-end speech model capable of seamlessly generating and processing high-quality roleplay audio.
  • They are establishing a comprehensive pipeline to evaluate new speech model submissions against real-time performance benchmarks.
  • They will integrate their speech model into the Dippy app.
  • They are committed to pushing the boundaries of speech roleplay through iterative improvements and continuous data collection.

 

They recognise the complexity of developing a state-of-the-art speech model, so they are structuring the process into three distinct phases.

Phase 1:

  • They are launching a subnet with a robust pipeline for roleplay-specific TTS models, designed to interpret prompts for speaker identity and stylistic speech descriptions.
  • They are introducing an infinitely scalable synthetic speech data pipeline.
  • They are implementing a public model leaderboard, ranking submissions based on core evaluation metrics.
  • They are incorporating Human Likeness Score and Word Error Rate as live evaluation criteria for continuous model assessment.

Phase 2:

  • They will refine their TTS models to produce more creatively expressive and highly human-like speech outputs.
  • They will showcase the highest-scoring models, making them publicly accessible through their front-end interface.

Phase 3:

  • They aim to advance toward an end-to-end speech model capable of seamlessly generating and processing high-quality roleplay audio.
  • They are establishing a comprehensive pipeline to evaluate new speech model submissions against real-time performance benchmarks.
  • They will integrate their speech model into the Dippy app.
  • They are committed to pushing the boundaries of speech roleplay through iterative improvements and continuous data collection.

 

MEDIA

Huge thanks to Keith Singery (aka Bittensor Guru) for all of his fantastic work in the Bittensor community. Make sure to check out his other video/audio interviews by clicking HERE.

Angad and Akshat from the Impel team join Keith to discuss their launch on Bittensor’s Subnet 11, aimed at incentivizing the decentralized creation of roleplay models for their app Dippy.ai. Learn about the team, their impressive backgrounds, and their ambitious goal of becoming the open-source leaders in roleplaying LLMs.

Angad and Akshat join the pod for the second time to talk evolution of Dippy.ai and how they are using multiple subnets and integration within Bittensor’s network to further the reach and capabilities of their viral roleplaying app. With a successful subnet (S11) and second subnet (S58) launched to add voice to their offering, this team is becoming a major force both in and outside of Bittensor.

NEWS

Announcements

MORE INFO

Useful Links

View other subnets

Back to the Directory

Keep ahead of the Bittensor exponential development curve…

Subnet Alpha is an informational platform for Bittensor Subnets.

This site is not affiliated with the Opentensor Foundation or TaoStats.

The content provided on this website is for informational purposes only. We make no guarantees regarding the accuracy or currency of the information at any given time.

Subnet Alpha is created and maintained by The Realistic Trader. If you have any suggestions or encounter any issues, please contact us at [email protected].

Copyright 2024