Inspiration
ReconeroRMS was inspired by a problem we kept hearing about from law enforcement: many police officers still rely on very traditional records management systems for handling cases and evidence. These systems often feel rigid, outdated, and difficult to work with, especially in situations where speed and clarity matter most.
We also learned that writing reports can take officers hours, particularly when they are newer to the job and still developing confidence in documentation and case preparation. That creates a major time burden and can pull attention away from the actual investigative work.
After hearing a talk from an officer at our local police department, we decided to learn more directly. We interviewed them about their existing RMS workflows, the friction points they deal with, and where current systems fall short in the field and in the office.
That process shaped our core belief: law enforcement could be far more efficient if their tools did not feel like they were working against them. We wanted to build a system that works with officers, helping them organize information, reduce repetitive work, and make better use of the data already in front of them.
What it does
Reconero RMS is an AI-powered records management and investigative case platform that helps users organize complex investigations as interactive relationship graphs. Rather than storing information in disconnected files, the system lets users connect people, vehicles, locations, evidence, notes, phone numbers, testimony, and custom entities in a shared visual workspace.
The platform supports collaborative case management with department-based organization, case-level access control, invite flows, and persistent shared workspaces. Users can upload and manage images, documents, and video evidence, attach them directly to case entities, and review investigations through graph, list, and map-based views.
A major differentiator is the AI layer built into the workflow:
- AI search helps users find relevant cases and entities across records, including semantic and deep-search workflows for narrative-style queries.
- AI analytics converts natural-language questions into structured analytics queries, then returns charts, maps, and trend views that help surface patterns in the data.
- AI case assistance can summarize nodes, generate case reports, answer questions about a case, highlight related entities, and assist investigators through conversational chat.
- AI media analysis supports evidence review by analyzing uploaded photos and generating structured information from media, as well as transcribing and summarizing uploaded video evidence.
- Live AI interaction gives users a more dynamic assistant experience for exploring and reasoning through cases in real time.
- Alyx Mobile App supports a mobile evidence capture workflow centered around the device camera, allowing field users to quickly capture photos and bring evidence into the case system or ask Alyx to identify objects found in other cases. This extends Reconero RMS beyond desktop case review and supports faster evidence collection in real investigative scenarios.
How we built it
Technologies Used
- Frontend: React, TypeScript, Vite
- Backend: Node.js, TypeScript, Fastify
- Database: PostgreSQL with Prisma
- Authentication: JWT, bcrypt
- Visualization and Mapping: Vega, Vega-Lite, MapLibre GL, Google Maps
- AI Integrations: Google Gemini, Google Vertex AI
- Utilities: Zod, exifr
Other Data Sources Used
- User-created data: case records, entities, relationships, and uploaded evidence
- Seeded/demo data: built-in demo case data and mock analytics datasets
- External APIs and services:
- NHTSA vPIC API for VIN decoding
- Google Maps services for map rendering
- Rest Countries API for country and phone metadata
Challenges we ran into
Live updates. Building a system where case changes feel immediate and collaborative introduced a lot of technical complexity. Keeping shared workspaces synchronized in real time, especially while supporting AI-driven interactions and case editing, was one of the harder engineering problems we faced. For live updates, we implemented a more structured real-time sync approach so shared case activity could stay consistent across users.
Cloud deployment. Getting the platform deployed in a stable way required us to think through hosting, environment configuration, backend services, media handling, and making sure the full stack worked reliably outside of local development. Moving from a working prototype to a deployed experience was a much bigger step than it first appeared. We addressed these challenges by tightening the architecture between the frontend and backend and focusing on reliability over shortcuts.
Project Type. Now it is obvious that our project is a Live Agent, but at some point we had to decide whether we wanted to focus on the workflow by doing UI navigator, or if we wanted to opt more for a Creative Storyteller that would write reports with image generation. It ended up making more sense to have a Live Agent, as it was the most useful one for law enforcement.
Accomplishments that we're proud of
Live workspace understanding. One of the things we are most proud of is building a system where the AI can operate within the context of a live investigative workspace. Instead of treating a case like disconnected text, ReconeroRMS gives Alyx awareness of entities, relationships, evidence, and evolving case structure 24/7 without any resets. Multiple users can also collaborate on the same case whilst using Alyx.
Going beyond "See, Hear, Speak". We did not want the AI experience to stop at basic multimodal input and output. We pushed the project further by making AI useful inside actual investigative workflows, including organizing workspaces, moving nodes, creating/deleting connections, and case search & analytics.
A production-ready project. We are proud that ReconeroRMS goes beyond a narrow demo and functions as a more complete platform. Departments, analytics, search, collaboration, permissions, media handling, and AI-assisted workflows all work together as part of a broader product experience. Did we mention the collaboration feature?
Mobile app integration. We are also proud of extending the platform beyond desktop workflows. Integrating a mobile experience supports faster evidence capture and helps move the product closer to real field use, where officers need tools that work wherever they are.
What we learned
We enjoyed the process and had fun building it. As developers, one of the biggest takeaways was that building something ambitious is a lot more sustainable when the team is genuinely enjoying the work. We all love to build, and that kept us going for this hackathon!
Impact matters just as much as innovation. We learned that it is not enough to build something technically impressive. We had to stay grounded in the real problem we were trying to solve for investigators and records workflows. Starting with the problem, rather than the technology, made the product much more focused and useful. Unlike most projects, we chose to interview our target audience (police officers) before building.
Collaboration is challenging, but worth it. Working as a team was not always easy, especially when balancing ideas, priorities, and implementation details. But as developers, we saw firsthand that collaboration made the product stronger. Different perspectives helped us catch gaps earlier, improve decisions, and build something more complete than any one of us could have built alone.
What's next for ReconeroRMS
Live body camera integration with Alyx. We plan to connect Alyx directly to police body cameras in real time so it can assist officers in the field as events unfold. By combining live video/audio context with officer metadata such as location, Alyx could provide faster, more relevant support during active investigations and response situations.
Expanded automation across case workflows. We want to increase automation throughout the platform so that fields, entities, and case details can auto-populate when enough supporting information is available. This would reduce repetitive manual data entry, improve consistency, and help officers spend less time on paperwork.
Court case preparation tools. We plan to extend Alyx beyond investigation support into courtroom preparation. This includes helping officers organize evidence, review timelines, identify weak points or missing documentation, and prepare to present testimony more clearly and confidently in court.
Lastly, thank you Google for giving us the opportunity to build something so cool :)
Log in or sign up for Devpost to join the conversation.