Question-first audio for data teams
Your data catalog,
as a podcast
Start with a question, connect your catalog, and let two AI hosts turn the answer into something your team will actually hear.
No signup required · 30 seconds to hear it
Why a podcast?
Your warehouse has hundreds of tables. Nobody reads the docs. But everyone listens to podcasts.
Passive consumption
Listen while commuting, coding, or doing dishes. No screen required.
Issues you'd miss
AI hosts flag failing tests, stale tables, PII columns, and missing owners.
Click to explore
Hear something interesting? Click the segment to see columns, tests, and lineage in real-time.
How it works
Connect
Point at OpenMetadata or dbt
Analyze
AI scans schemas for health, lineage risks, and indexer lag
Listen & share
Stream episodes or share MP3s via Slack
What you get per episode
Plans
Start free. Upgrade when your team needs scheduled episodes.
Free
$0
- ✓ Unlimited one-off episodes
- ✓ All data sources
- ✓ MP3 download & sharing
- ✓ Interactive drill-down
Team
$29/mo
- ✓ Everything in Free
- ✓ Scheduled daily/weekly episodes
- ✓ On-chain health minting (Initia)
- ✓ Private team RSS feeds
- ✓ Slack/webhook notifications
- ✓ Historical comparison
Questions
Is my data sent to your servers?+
Your credentials are stored server-side in an encrypted session that expires after 1 hour. We fetch metadata from your catalog to generate the episode, but we don't store your raw data. Audio and episode data are cached temporarily for sharing (24h for free, 7 days for Pro).
Do I need an ElevenLabs account?+
For the demo, no. To generate episodes from your own data, you need an ElevenLabs API key (Starter plan at $5/mo is recommended for full API access).
What data sources are supported?+
We support OpenMetadata, dbt (Cloud & Local), The Graph (any subgraph), and Dune Analytics. We pull metadata, quality tests, lineage, and indexer status to build your health profile.
How does the Initia integration work?+
For protocol teams, every episode can be minted as a health attestation to the Initia appchain. This creates a verifiable history of data quality that can be shared with your community or DAO.
How long does generation take?+
30-60 seconds depending on schema size. The AI analyzes your metadata, generates a two-host script, then synthesizes each segment as audio. You see real-time progress with segment-by-segment updates.
Can I self-host DataBard?+
Yes. DataBard is a Next.js app that runs on any Node.js server. Clone the repo, set your API keys, and deploy. The file-backed store works on any persistent server. No database required.
What are the two AI hosts?+
Alex is the enthusiastic data advocate who highlights what's working well. Morgan is the skeptical quality auditor who flags risks, failing tests, and governance gaps. Together they create a balanced, engaging walkthrough.
What's the visual report feature?+
Click ‘📊 Report’ on any episode to download a 3-slide PDF health report — overview, critical tables & action items, and lineage & ownership. Generated server-side using Paper-inspired design, no external tools required.
How does the free tier work?+
Unlimited one-off episodes, all data sources, MP3 download, sharing, and interactive drill-down. Pro adds scheduled episodes, private RSS feeds, Slack notifications, and historical comparison.