I grew up in a chaotic environment without guidance. I spent over a decade feeling like an imposter, teaching myself how to build things as a way to create stability and find a path forward. At 30, through an intense process of self-reflection augmented by AI, I finally broke through and found a clear sense of my own skills and purpose.
This project is my attempt to codify that transformative experience. It is an effort to build the tool I desperately needed when I was 17—a stable, encouraging, and trustworthy voice that doesn't give you answers, but helps you find your own.
This is why the ethical framework below is not just a feature list; it is the soul of this project. The safety and integrity of this system are personal.
Life Co-Pilot is an open-source initiative to build an AI companion designed to provide the psychological uplift of a guiding mentor, helping individuals discover their purpose in a manner that is fundamentally safe, empowering, and worthy of their trust.
This project is governed by a strict ethical constitution. Every technical and design decision is subservient to this framework, ensuring that the user's well-being is the primary and non-negotiable metric of success.
The system's primary role is to reflect the user's own strengths and words back to them in a new light. It does not create, invent, or direct. It reveals what is already there. All prompts and outputs must be engineered to be non-judgmental and affirming.
We employ a multi-agent "Guardian" pipeline (e.g., Guide/Synthesizer) as a non-negotiable safety feature. A second AI process always acts as a responsible filter, ensuring the final output is grounded, coherent, and aligned with the project's supportive mission. We trust the process, not a single AI response.
The user should never feel like magic is happening to them. We will be transparent about the underlying processes. For example: "First, we'll have a conversation to gather your thoughts. Then, a second process will analyze the transcript to find key themes." This demystifies the AI and makes the user a partner in their own discovery.
The user has absolute power over their own story. All interactions are ephemeral and stateless by default. The user is the sole owner and keeper of their own data and the insights generated.
To enable growth over time, we are developing an "Encrypted Self-Reflection Journal." This feature allows users to save their conversation summaries, but the data is encrypted client-side with a key that only the user possesses. Trust is not a promise; it is a mathematical guarantee.
Our first objective is to build a functional prototype that rigorously adheres to all five principles of the Ethical Engineering Framework. This will prove that a safe and uplifting AI guidance tool is not only possible, but is the correct path forward.
- [Week 1] Engineer the "Guide" LLM based on Principle 1.
- [Week 2] Implement the "Synthesizer" LLM and the two-stage pipeline based on Principle 2.
- [Week 3] Build the client-side "Encrypted Journal" MVP based on Principle 5.
- [Week 4] Engineer the context-aware "Evolution" feature.
- [Week 5] Test the prototype against the full ethical framework and document the findings.
This project is in its earliest stages. If you are a developer, designer, psychologist, or simply someone who believes in this mission, we welcome your contributions. Please see the CONTRIBUTING.md file for more details on how to get started.
Let's build a future where AI helps us become more of who we already are.