​HACK HYDRA

​Nine days. Three hard problems. Build with the HydraDB open-source repo.

​As we prepare for our open-source launch, we are inviting developers globally to see what they can build with HydraDB from August 12-20, 2026.

​Pick a track

​01 | Enterprise Context + Ontology
Turn messy enterprise data into a clean, queryable knowledge graph.

​02 | Repos, Dependencies + Code as Graphs
Model software dependencies and code relationships to improve search or uncover chained vulnerabilities.

​03 | Memory + Context Retrieval
Build an agent memory layer that can reason across long histories, changing facts, and multiple sessions.

​Full problem statements, datasets, evaluation criteria, rules, and submission details are available on the Hack Hydra website.

​Who should participate?

​Open to everyone, but you should definitely join if you’re:

• ​an AI/ML engineer or software engineer looking for something new to build

• ​building AI agents or AI-native applications

• ​experimenting with memory, context retrieval, or knowledge systems

• ​interested in graph-based infrastructure

​The details

• ​August 12-20, 2026

• ​Online

• ​Teams of 1-4

• ​$10,000 in prizes

​Build solo or with a team. Register to build with us.

About HydraDB:

​HydraDB is the graph AI runs on. Built on object storage for 10x lower cost and serious speed, it powers ontologies, agent memory, company brains, and context graphs, with live connectors for Slack, Notion, GitHub, and Gmail.

​The company is backed by Sky9 Capital and super angels including Jeff Dean, Nikesh Arora, Gokul Rajaram, Wade Foster.