SkyJac

• 

  Find Partners Tab

• 

  Climate Change

• 

  Gaps

• 

  Coalition

• 

  Organisations

Inspiration

Multilateralism is weakening. International institutions are under strain, trust between nations is eroding, and global challenges like climate change, food insecurity, and refugee crises go unaddressed. Not because solutions don't exist, but because the right organisations never find each other. SkyJac exists to fix that coordination failure.

What it does

SkyJac is an agentic coalition partner matching platform. You select a global challenge, and SkyJac surfaces every organisation aligned on that cause, showing mission, capacity, region, and a focus score. You can then propose a coalition, select members, and form it in one flow. A live Gap Analyst tab shows which challenges have zero coalition coverage, with a direct "Start Coalition" action to close each gap.

How we built it

Built entirely on Jac and the Jaseci fullstack stack: One .jac file compiles to both a FastAPI backend and a React frontend simultaneously. No separate codebases, no HTTP glue to write. Jac's graph-native model maps perfectly to coalition relationships: orgs are nodes, shared causes are edges, and walkers traverse the graph to find alignment. All data is seeded locally for the demo with 12 real-world organisations across 8 global challenges.

Challenges we ran into

Jac's in operator compiles to JavaScript's in, which checks object keys rather than array values. Every list membership check silently returned false until we replaced them all with explicit list comprehensions. Learning this quirk cost real time but taught us exactly where Jac's Python-to-JS boundary lives.

Accomplishments that we're proud of

A fully working agentic coalition platform in a single .jac file, built during the hackathon while learning Jac from scratch. The Gap Analyst closing in real time is the moment the whole thing clicks. We can propose a coalition for a gap challenge, submit it, and once it confirms, the Gap closes.

What we learned

Jac is genuinely a different programming model. Writing the full stack as one program, where the compiler decides what runs in the browser and what runs on the server, changes how you think about data flow. The by llm() pattern for replacing function bodies with AI is the most elegant AI integration we've seen.

What's next for SkyJac

SkyJac is designed as a 24-module platform. The remaining modules include the Overlap Detector, Trust Signal Scorer, Momentum Tracker, Policy Window Alerts, Pledge Engine, Resource Pooling Map, Action Plan Generator via by llm(), Coalition Health Monitor, Success Story Surfacer, Narrative Reframer, Public Dashboard, UN/OECD data sync, Impact Reporter, and Simulation Mode. The architecture is modular by design. Each module can be inserted and removed without affecting the whole thing.

Built With

• bun
• fastapi
• jac
• jaseci
• react