GoalKeeper

The description states that GoalKeeper is an AI accountability app for focused work. It allows users to:

• Define a goal.
• Choose visible behaviors they want help avoiding (e.g., phone use, leaving workspace).
• Set a focus timer and session contract.
• Receive check-ins only when sustained evidence accumulates.
• Respond to interventions via voice or text.
• End or override sessions.
• Review session history.

It uses:

• GPT-5.6 Luna for perception and reasoning.
• OpenCV for image capture.
• SQLite for local data storage.
• .NET 10 Blazor for UI.
• TTS and transcription for optional voice interaction.

The system is designed to avoid reacting to single ambiguous moments, instead accumulating evidence over time before intervening.

Inference GoalKeeper appears to be a prototype focused on AI-driven accountability in work environments. It is not described as a commercial product or platform with users beyond the developers themselves.

The description states that GoalKeeper was inspired by the lack of context-aware focus tools. The central claim is:

“Can an AI accountability partner accumulate uncertain evidence over time and intervene only when speaking is more helpful than staying silent?”

This positions GoalKeeper as a tool that avoids over-interrupting users while still providing support.

Inference The positioning reflects a shift from simple distraction-blocking tools to ones that understand context. However, the description does not indicate any evolution in claims beyond this core idea — no indication of how it might scale or adapt to different use cases.

The description does not state a specific customer segment or ideal customer profile (ICP). It describes the app as being for users who want to focus, but does not identify:

• Industry
• Role (e.g., student, professional)
• Work environment
• Pain points beyond distraction

Inference The target audience seems to be individuals seeking better focus tools — likely professionals or students. But no explicit ICP is defined.

There is no evidence of a business model or pricing structure in the description. The project is presented as a hackathon submission with no mention of monetization, subscriptions, or paid features.

Inference The app appears to be non-commercial at this stage, likely intended for demonstration or internal use only.

The system is built using:

• .NET 10 Blazor
• C#, ASP.NET
• Entity Framework
• GPT-4o and GPT-5.6 Luna (for perception and reasoning)
• OpenCV
• SQLite
• TTS and transcription models
• xUnit for testing

Key technical features include:

• Local storage of session data.
• Isolated roles for perception, reasoning, recovery.
• Timer control remains in .NET code, not AI.
• Camera preflight confirmation.
• Session contract validation.
• Handling of model failures or outdated responses.

Inference The architecture shows a strong emphasis on safety and user control. The separation between AI and state-changing logic suggests an intentional design to prevent unintended behavior.

There is no evidence of traction, revenue, or customer adoption beyond the hackathon submission. The description mentions:

• A recorded hosted run.
• Validation through 250-capture pipeline soak.
• 250-snapshot SQLite persistence test.
• Automated tests and concurrency handling.

However, these are not indicators of real-world usage or product maturity.

Inference The project is at a prototype stage. No evidence of user engagement, retention, or commercial viability exists.

The description states that existing focus tools measure time or block websites but do not understand context. GoalKeeper aims to differentiate itself by:

• Accumulating uncertain evidence.
• Intervening only when justified.
• Supporting rather than policing the user.

No mention of competitors or market positioning beyond this differentiation.

Inference GoalKeeper is positioned as a novel approach in the distraction management space, but no competitive analysis or market data is provided.

1. No commercial traction or revenue: The project is described only as a hackathon submission.
2. Unverified claims: All descriptions are self-reported and unverified.
3. Prototype-only status: No evidence of product-market fit, user feedback, or scalability.
4. AI dependency risks: Reliance on GPT-5.6 Luna for perception and reasoning introduces potential instability or bias.
5. Privacy concerns: Use of room-facing camera with local storage raises privacy questions, though mitigation strategies are described.

Inference The project is experimental and not yet a viable commercial product. Risks include lack of user validation, technical fragility, and unclear path to monetization.

1. What specific use cases or industries were you targeting with this prototype?
2. How did you validate the effectiveness of interventions (e.g., did users find them helpful)?
3. Are there any plans for user testing beyond the hackathon?
4. What are your thoughts on privacy implications and how do you plan to address them at scale?
5. Have you considered how to handle edge cases like camera failure or model hallucinations?
6. Is there a roadmap for moving from prototype to a commercial product?