LifeOps

Inspiration

Moving to a new country is one of the most stressful transitions a human can undergo, especially when you don't speak the native language. The cognitive load of navigating foreign bureaucracy, employment laws, and cultural shifts often leads to "analysis paralysis." We wanted to build a co-pilot that doesn't just answer questions, but proactively maps out the journey, remembers the context, and transforms fragmented concerns into a clear, actionable signal.

What it does

LifeOps is a premium AI navigator for immigrants. Users can speak their concerns or updates naturally, even with varied languages, accents or emotional tones. The agent extracts "Context Facts" (e.g., arrival date, employment status, visa types) and builds a long-term memory graph in Neo4j. It then autonomously infers necessary tasks, like applying for a California State ID or drafting a W-4, and grounds them in real-world data by finding official government resources for the user.

How we built it

Long-Term Memory: Utilized Neo4j to maintain a User Context Graph, allowing the agent to reason across past conversations. Audio Intelligence: Integrated the Modulate API (Velma STT) for high-fidelity transcription, optimized for handling emotional tone and speaker diarization. Agentic Grounding: Used the Yutori API to autonomously research official government requirements and verify links for every inferred task.

Challenges we ran into

Our primary challenge was building a robust voice-to-reasoning pipeline. Ensuring the agent's inferences remained grounded in official sources while handling the high information density of a multi-node life journey required significant tuning.

Accomplishments that we're proud of

We are incredibly proud of the end-to-end Context Ingestion pipeline. Seeing a user's natural, spoken concerns transform into structured Neo4j nodes and then automatically trigger Yutori-sourced government links felt like magic. We also succeeded in creating a distinctive, premium visual identity that breaks away from generic AI desktop tools.

What we learned

We learned the critical importance of Grounding in autonomous agents, reasoning is only useful if it's connected to current, official documentation. We also gained deep experience in modeling personal life events as graph entities and leveraging emotional signals from speech to better understand user intent.

What's next for LifeOps

The next phase for LifeOps is Autonomous Execution. We want to move beyond finding the links to helping users auto-fill complex immigration forms and schedule biometric appointments directly. We also plan to integrate localized community context, helping global citizens navigate not just the legal "next steps," but the social and cultural ones too.

GitHub: https://github.com/pkhiani/lifeops-agent

Built With

• modulate
• neo4j
• render
• yutori