Aletheia - Clarity in Care

https://github.com/siddharthsiva/aletheia/tree/main

Inspiration

• Healthcare is often riddled with hidden costs, fine print, and vague answers
• Many patients face surprise bills, misunderstood meds, or denied claims
• We wanted a system where clarity is the norm and profit-driven ambiguity is impossible

What it does

• Parses and uploads medical documents to build a living health profile
• Logs pills, sends reminders, and gives next-step advice if doses are missed (especially for urgent conditions)
• Identifies any pill from a photo and explains its real effects on you and gives you simple ingredient breakdowns
• Scores insurance providers using a trust index based on controversies, verified reviews, and trusted sources. Also recommends higher rated providers based on your context
• Provides a transparent conversational health consultant that connects directly with a healthcare consultant via Slack API

During this entire process, the AI's thought process and returns are completely transparent to the user - leaving no hidden AI reasoning.

How we built it

• Designed a multi-agent system: each agent focuses on one task (parsing, pill ID, insurance scoring, advice).
• Agents share a global context - your secure medical history and background
• Agents query each other or trusted APIs (FDA data, Slack for doctor contact)
• Combined this backend w/ intuitive UI for doc uploads, pill logs, and AI chat

Challenges we ran into

• Coordinating specialized agents while avoiding overlaps and contradictions
• Balancing transparency with data privacy and security
• Ensuring the information provided by AI were ACCURATE and NOT misleading through continuous refinement of filtering process and query functions

Accomplishments that we're proud of

• Built a working hospital-like AI network that understands real documents and pill images
• Was able to successfully pivot last second and still retain the impact and features
• Made all AI thoughts and decisions open to the user - ensuring we held up transparency ourselves

What we learned

• Multi-agent orchestration requires clear domain boundaries and inter-agent trust
• Personal context is critical for meaningful healthcare advice
• Transparency builds trust - users want to see exactly how AI thinks
• Bridging AI with human experts to ensure AI is kept in check and safe

What's next for Aletheia - Clarity in Care

Deepen integrations with real healthcare providers and pharmacists.

Expand the trusted data pipeline: more medical sources, insurance databases, verified reviews.

Refine the user interface for even simpler clarity and control.

Evolve into the gold standard for transparent, personalized, life-first care.

Built With

• letta
• pil
• python
• streamlit