Pirates of theCoral-bean
Hoist the Jolly Roger and set sail on the high seas of data! Build agents that query any API, database, or file as SQL, powered by Coral. No ETL, no warehouse, no glue code. Just you, your crew, and the open ocean of data.
-- Find root cause across 3 tools in one query
SELECT g.title, s.error_message, sl.text
FROM github.pull_requests g
JOIN sentry.issues s
ON s.first_seen >= g.merged_at
JOIN slack.messages sl
ON sl.channel = '#incidents'
WHERE s.level = 'fatal'
ORDER BY s.first_seen DESC;
About the Voyage
In a world where AI agents need data from every corner of the seven seas, there's only one tool that lets them query it all as SQL.
Coral is an open-source data retrieval layer for agents. It lets them query any API, database, or file as SQL tables. Coral handles auth, pagination, rate limits, cross-source joins, schema learning, and caching.
No ETL. No warehouse. No glue code. Everything is local. Credentials, data, and usage history never leave your machine.
"Not all treasure is silver and gold, mate. Sometimes it's a perfectly joined SQL query."
Query Anything as SQL
Any API, database, or file becomes a SQL table. No custom integrations needed.
Cross-Source Joins
Join GitHub + Slack + Sentry in a single query. Coral handles auth, pagination, and rate limits.
100% Local
Credentials, data, and usage history never leave your machine. No ETL, no warehouse.
CLI or MCP
Run Coral from the command line or through MCP. Schema learning and caching built in.
The Treasure Map
Two paths to glory on the high seas of data. Choose your adventure.
Track 1: Build an Agent
Captain your own data vessel
Build an agent for a real use case that retrieves data across multiple sources using Coral. Describe how you did it and what makes it great, ideally showcasing Coral's power.
Example Voyages
Track 2: Build a Source Connector
Expand the Coral reef
Connect any API you wish Coral supported. Build a custom Coral source for an API that doesn't exist yet. This is a high-value contribution that grows the sources ecosystem.
How to get started
The Treasure Chest
$10,000+ in gold doubloons and bounties for the bravest pirates on the data seas.
Captain's Bounty
All team members will receive the prize for their placement.
Track 1 Winners
Track 2 Winners
Early Bird Swag: Top 10 Social Sharers
Register and share on social media (LinkedIn / X) tagging Coral.
⚔️ Special Bounties
Beyond the main prizes, earn extra treasure through these challenges.
Discord Showcase + Social Post
Share your project in Discord #how-i-coral + at least one social platform with screenshots.
Build "Wanted" Source Specs
We'll share a list of 10 "wanted" new sources (Hubspot, Asana, Zendesk, Airtable, etc.).
"Before & After" Demo Videos
Record a max 3-min video: building without Coral vs. with Coral.
End-to-End "How to Build X" Guides
Write a 2-3 page reproducible guide, e.g. "I built a dashboard with Claude Code + Coral."
Recruit Yer Crew: Referral Bounty
Refer 5 people who register AND submit a project.
Ready to set sail? Register and start building!
Sponsor
Coral is an open-source data retrieval layer for agents that lets them query any API, database, or file as SQL tables. Coral handles auth, pagination, rate limits, cross-source joins, schema learning, and caching. No ETL, no warehouse, no glue code.
You can run Coral from the CLI or through MCP. Everything is local. Credentials, data, and usage history never leave your machine.
How to use it? For hackathon participants, Coral provides a unified SQL interface to query data from multiple sources. Instead of writing custom API integrations, pagination logic, and auth flows for each data source, you write SQL queries and Coral handles the rest. Focus on building your agent's logic and impact while Coral handles the data plumbing.
Judging Criteria
🏴☠️ Potential Impact
How effectively does the project address a meaningful problem using Coral's data retrieval capabilities?
⚓ Creativity & Originality
How unique is the idea? How creatively is Coral used to solve problems across multiple data sources?
🗺️ Learning & Growth
The learning curve tackled during the hackathon, especially for first-time Coral users.
⚔️ Technical Implementation
How well was the idea executed technically? Quality of integration with Coral and SQL queries.
🎨 Aesthetics & UX
How intuitive and user-friendly is the final experience? Frontend, CLI, or agent interface.
🪸 Best Use of Coral
How effectively does the project leverage Coral's SQL interface, cross-source joins, and caching?