Project Zeta Funding Pitch

Project Zeta has now full outlined its future goals for the project. You can watch the video or read below.

What is Project Zeta:

Project Zeta aims to be an alternative to human interaction for people who may not be in a situation to meet new people, or just need someone to talk to when nobody else is available. Project Zeta currently has a functioning prototype for both Android, PC, PC VR and Meta Quest devices. This project will be available to people who are over 18 due to the nature of subject matter the AI is able to talk about.

Project Zeta’s Scope:

Project Zeta has achieved the initial scope of being a companion that can successfully hold back and forth conversations with high quality generated speech. The next steps are to improve the general server infrastructure and speed up responses to the end user for more fluid and faster conversations. This prototype has been put together with a budget of $0 and a lot of hard work in a short amount of time. There’s plans to add activities and different locations that people can take their AI companions to. Improving the visual style of the avatars with higher quality assets will also be a long term priority in order to improve immersion once there's enough funding to do so. The end goal is for Project Zeta to be able to satisfy the social needs of people in a way that is more immersive and nuanced than existing AI companion services on the market today and in a way that is fully unrestricted to the end user. You should be able to do as you wish, and talk about any subject you want to with your companion. The technology and how we handle data.: The full feature set of Project Zeta uses the following technologies to come together. We use pre-trained AI speech recognition that runs fully and completely on your device. Despite being able to use your own voice to talk and converse, we never transmit or hear your audio. We don’t have access to this data. Your device will only ever send an anonymous SSL encrypted text transcript to the AI, and only after you press a button with intent to speak to the AI. In order to generate responses we use fully open source and self managed pre-trained AI Models. In order to ensure users can talk about a wide range of topics without fear of censorship, we don't use closed source corporately owned AI’s with moderated usage such as ChatGPT. We also do not use any user data to train the AI models we use. In order to generate the voice for each character we use a high end AI voice service. This service only ever gets to store the AI’s own responses in a disjointed and incoherent manner and can barely be used to reconstruct its own side of conversations. At the fullest extent, we get a vague idea of what the AI is saying back to the end user if we consciously decide to moderate the AI’s own behavior. We have no context other than the pure audio files being generated for each phrase. In order to save on processing costs we automatically re-use generated audio for each voice if a phrase is common enough to be used again. This audio data is stitched together to create a full audible sentence. Project Zeta does not store any direct conversation history or transcripts on any of the server infrastructure used for generating dialogue. All data submitted to the server for processing is stored and persisted on the end user’s device and is only ever used to generate the AI’s next response as needed. The user has full control of their own data at all times and is responsible for keeping it safe from data loss. Project

Zeta’s Financial Plan:

Project Zeta charges a monthly subscription to access the full features of the application. This money is used to maintain server costs, and to pay for further development and maintenance of services. Cloud services are currently used to generate AI dialogue and audio, and these services have an inherent cost that needs to be paid for the service to be maintained.

Our goal is to be funded by the community, and directly by the end user. For as long as the project is able to receive funding, the project will continue to move forward with updates and evolve based on user feedback.