01 / The challenge
Build a working agent, or a group of agents, with FalkorDB as its context layer. The graph has to shape what the agent does, not sit on the side.
The problem
Agents see pieces, not the whole picture. They lose track of what happened in earlier sessions, miss how people, systems, and records relate, and can't explain why they made a choice.
The shift
Put the agent's context in a graph. Data, memory, and company knowledge become nodes and edges the agent can follow, so it works from real relationships instead of guessing.
The brief
Build an agent that acts on connected data, remembers what it learns, and can show how it got to each answer.
Work the graph has to do
Pick a task where the answer depends on how things connect, and build the agent that handles it.
Enter on your own or with up to three others. Register once, either way.
Free to enter, online from anywhere, and no graph database experience needed.
Agents that act on connected data, agent memory and coordination, or company brain. Enter the one your idea fits.
02 / Tracks
Build for one, or for all three at once. Each track has its own prize.
Track 01
Build an agent that uses the graph to decide and act. It could investigate fraud, triage security findings, or work through a codebase.
Key focus
Track 02
Build agents that remember across sessions and share state with each other. Memory should persist, update over time, and be available to every agent that needs it.
Key focus
Track 03
Build an agent that understands how a company works. Connect people, teams, projects, documents, and decisions in one graph, so the agent can answer questions and get work done across all of it.
Key focus
Prize pool
$20,000
One prize for each track, plus two side prizes. Every member of a winning team gets the prize.
Agents That Act on Connected Data
For the best agent that uses a graph to make decisions and take action.
Agent Memory and Coordination
For the best agents that remember across sessions and share what they know.
Company Brain
For the best agent that understands how a company works and gets work done.
Side prize
AirPods 5
1 winner
Write a blog post about your project. The best post wins a pair of AirPods 5.
Side prize
T-shirt
Top 50 projects
Every member of the top 50 teams gets a t-shirt, shipped to their address.
03 / Why graphs
The hard questions agents get asked don't live in one record. The answer depends on how several things relate.
Which accounts are linked to this flagged transaction?
What did this customer ask for last month, and was it resolved?
Who on the team has worked on this system before?
What breaks if this service goes down?
A graph database stores those relationships directly, so the agent can follow them in one query instead of stitching records together on every step.
04 / About the sponsor
An open-source graph database built for AI agents.
Traversals are fast enough to run on every step of an agent loop, and one instance can hold thousands of isolated graphs, so each user, agent, or tenant gets its own. Run it locally with Docker.
docker pull falkordb/falkordb:latestOfficial clients
Cypher
who knows about this system
MATCH (p:Person)-[:WORKED_ON]->(:Project)-[:USES]->(s:System {name: $system})
RETURN p.name AS person, count(*) AS projects
ORDER BY projects DESC
LIMIT 5A graph your agent can read and write on every step, and still find on the next session.
Use FalkorDB through the framework or protocol you already build with.
Leverage multigraph topology to create isolated graphs in a single instance, without running separate deployments.
Create nodes and relationships, match patterns, follow paths, and return what your agent needs.
Traversals your agent can call instead of writing them itself.
Use FalkorDB Browser to run queries, inspect the data model, and see how records connect.
05 / FAQ
From October 15, 12:01 AM IST to October 18, 11:59 PM IST. Submissions close at 11:59 PM IST on October 18.
The hackathon is open to developers worldwide. You can participate solo or with a team of up to four people.
No. The getting started guide and the kickoff session cover the basics you need to begin.
Yes. FalkorDB must be the primary graph database, and it must support a central part of the product.
Yes. You can enter one, two or all three, and each track has its own prize. As a guide: an agent that acts on data fits Track 01, one that remembers or coordinates fits Track 02, and one that works with company knowledge fits Track 03.
Right here on WeMakeDevs. Sign in, or create a free account, then hit Register now on this page. Your team and your submission are managed from the same panel.
Yes. You may use any language, framework, frontend, or deployment platform that works with your project.
You may reuse general libraries, templates, and infrastructure. The submission and its main FalkorDB implementation must be new work completed during the event.
Yes. You remain responsible for the code, architecture, security, and correctness of the project, and you must be able to explain what was built and why.
A live deployment is preferred. Projects that cannot be deployed must include complete local setup instructions and a working demo video.