🎙️ DataBard

Turn your onchain data into AI podcasts and music anthems — then mint them on Solana.

Connect your Dune queries, subgraph, or data warehouse. Two AI hosts analyze the data, narrate the health story, and generate a shareable episode or a full Data Anthem (music track). Mint the result on Solana under your .sol identity.

▶ Listen to a demo episode — no signup required · 🎵 Hear a demo anthem

Hackathon reviewers:

Colosseum Frontier / 100xDevs Track → Live product: databard.thisyearnofear.com · Solana minting · ElevenLabs Music API · SNS .sol identity
Palm USD x Superteam UAE → Native PUSD payment for Pro subscriptions · SPL token transfer on Solana · non-freezable stablecoin checkout · docs/PALM_USD_INTEGRATION.md
SNS Identity Track → .sol domain resolved on wallet connect, stored in on-chain memo at mint time

The Problem

Data catalogs have hundreds of tables. Nobody reads the docs. Quality tests fail silently. New team members take weeks to understand the warehouse. The metadata is there — but nobody consumes it.

The Solution

DataBard generates podcast-style audio episodes from your OpenMetadata instance. Two AI hosts — Alex (the enthusiast) and Morgan (the skeptic) — have a natural conversation about your schemas, surfacing insights that matter:

🔴 "orders has 3 failing tests and 8 downstream dependents — if this breaks, it cascades"
🔒 "Governance flag: customers has PII columns — email, phone_number"
👤 "This table is owned by the analytics team. Hey analytics team, your tests are failing"
📊 "Health score 58 out of 100. Only 50% test coverage. Let's talk about what's broken."
🕐 "events hasn't been updated in 36 hours. Are the pipelines running?"

Click any segment to drill down into the actual columns, tests, lineage, and tags.

How It Uses OpenMetadata

DataBard deeply integrates with OpenMetadata's API surface:

API	What DataBard Does With It
Tables (columns, tags)	Discusses schema structure, column types, data modeling choices
Data Quality / Test Cases	Flags failures, computes health scores, identifies cascading risks
Lineage	Maps data flow, identifies hotspots, warns about upstream failures
Owners	Names who's responsible for broken tables
Profiler Data	Reports row counts, flags stale tables by freshness timestamps
Glossary Terms	Highlights business context baked into metadata
Classification (PII/Sensitive)	Governance alerts for columns with PII tags
Database Schemas	Schema descriptions, catalog browsing

The analysis layer computes health scores, critical table rankings (failing tests × downstream dependents), coverage gaps, and lineage hotspots before generating the script.

Architecture

┌─────────────────────────────────────────┐
│          Tier 1: First-Class            │
│  OpenMetadata · dbt · The Graph · Dune  │
│  (full lineage, PII, owners, tests)     │
├─────────────────────────────────────────┤
│     Tier 2: Coral Escape Hatch          │
│  Any source via SQL · cross-source joins│
│  (generic columns + data types)         │
└───────────────────┬─────────────────────┘
                    │
                    ▼
          ┌──────────────┐     ┌─────────────┐     ┌──────────┐
          │   Analysis   │────▶│   Script    │────▶│  Audio   │
          │ Health score │     │ LLM / tmpl  │     │ ElevenLabs│
          │ Critical tbl │     │ Two hosts   │     │ TTS + SFX │
          │ PII, owners  │     └─────────────┘     └──────────┘
          │ Lineage risk │                               │
          └──────────────┘                               ▼
                                                 ┌──────────────┐
                                                 │  Interactive │
                                                 │   Player     │
                                                 │ Drill-down   │
                                                 │ Share/RSS    │
                                                 └──────────────┘

Pipeline: Question → Metadata fetch (Tier 1 adapter or Coral SQL) → Schema analysis → Script generation (LLM or template) → Streaming audio synthesis → Interactive player with data drill-down

Features

Tier 1 adapters — Deep integration with OpenMetadata, dbt, The Graph, Dune (lineage, PII, owners, quality tests, profiler — zero extra dependencies)
Tier 2: Coral escape hatch — "Bring Your Own Source" via SQL for the long tail (Salesforce, Jira, Postgres, local files, cross-source joins)
Question-first analysis — Zerve-aligned workflow that starts from a research prompt, not a blank script
Research trail — each answer now includes a plan, evidence, and recommended next actions
Two-voice AI podcast — Alex (advocate) and Morgan (auditor) via ElevenLabs
LLM-powered scripts — GPT-4o-mini generates natural dialogue (template fallback)
Interactive drill-down — click any segment to see columns, tests, lineage, tags
Health scoring — 0-100 score based on test coverage, failures, documentation, freshness
Streaming synthesis — start listening while audio generates
Multi-platform sharing — WhatsApp, Telegram, Twitter, native share sheet, RSS feed
MP3 download — take episodes offline
Smart caching — SFX cached 30 days, speech 24hr, scripts 1hr, metadata 5-10min
Rate limiting — 5 episodes/hr per IP to protect API credits
OG images — dynamic social preview cards for shared episodes
Monetization-ready — Stripe checkout, Palm USD (Solana stablecoin), private RSS feeds, cron-ready regeneration

Quick Start

git clone https://github.com/thisyearnofear/databard.git
cd databard
npm install

cp .env.example .env
# Add your ELEVENLABS_API_KEY (Starter plan or higher recommended)
# Add OPENAI_API_KEY for LLM scripts (optional)
npm run dev

ElevenLabs Setup:

Sign up at elevenlabs.io
Upgrade to Starter plan ($5/month) for API access
Get your API key from Profile → API Keys
Add to .env: ELEVENLABS_API_KEY=sk_your_key_here

OpenMetadata setup:

curl -sL -o docker-compose-postgres.yml \
  https://github.com/open-metadata/OpenMetadata/releases/download/1.10.4-release/docker-compose-postgres.yml
docker compose -f docker-compose-postgres.yml up --detach

Open localhost:3000 → Connect → Select a schema → Listen.

OpenMetadata sandbox mode (no user token required)

If you want users to connect instantly from the frontend using Use Sandbox, configure these env vars on the server:

OM_SANDBOX_URL=https://sandbox.open-metadata.org
OM_SANDBOX_TOKEN=your_server_side_sandbox_bot_jwt
NEXT_PUBLIC_OM_SANDBOX_URL=https://sandbox.open-metadata.org

OM_SANDBOX_URL and OM_SANDBOX_TOKEN are used server-side by /api/connect when sandbox mode is selected.
NEXT_PUBLIC_OM_SANDBOX_URL is display-only in the UI.
Users can still switch to Connect Your Instance and provide their own URL/token.

Tech Stack

Layer	Technology
Web UI	Next.js 15, React 19, Tailwind CSS 4
Metadata (Tier 1)	OpenMetadata REST API, dbt manifest parsing, The Graph, Dune Analytics
Metadata (Tier 2)	Coral — "Bring Your Own Source" SQL layer for long-tail connectors + cross-source joins
AI Scripts	OpenAI-compatible API (GPT-4o-mini default)
Audio	ElevenLabs TTS (two voices) + Sound Effects
Caching	File-backed with TTL (no external dependencies)
Payments	Stripe Checkout, Palm USD (Solana SPL stablecoin)
Onchain	Solana (Memo Program + PDA registry), SNS `.sol` identity, episode minting, health alerts, leaderboard
Development	Kiro (spec-driven, see `.kiro/` directory)

Solana Integration

DataBard uses Solana not just for minting receipts, but as a genuine utility layer for data teams: queryable audit trails, health alerts, team accountability, gated access, and a public leaderboard.

Solana Utilities

Feature	API	Description
Palm USD payments	`POST /api/checkout/palmusd`	Pay for DataBard Pro ($29/mo) using Palm USD, a non-freezable Solana stablecoin. Server builds unsigned SPL transfer → wallet signs → `POST /api/checkout/palmusd/verify` confirms on-chain and activates Pro
On-chain audit trail	`POST /api/onchain/mint-solana`	Every episode mint writes a memo + PDA record: `(schema_fqn, health_score, timestamp, episode_id, wallet, .sol domain)` — permanently queryable via RPC
Health alert subscriptions	`POST /api/onchain/alerts`	Register a wallet + schema + threshold + webhook; `GET /api/onchain/check-alerts` fires Slack/webhook when health drops below threshold (cron-ready)
Public leaderboard	`GET /api/onchain/leaderboard`	Ranked protocols by latest health score, trend (↑↓→), mint count, wallet count — live at /leaderboard
Team history	`GET /api/onchain/team-history`	Cross-wallet mint history for a schema — shared ground truth for post-mortems and handoffs, surfaced in the EpisodePlayer "👥 Team" tab
Gated episode access	`GET /api/onchain/access`	Episode replay checks wallet ownership of the mint record; non-holders see a wallet-connect nudge
SNS `.sol` identity	`src/lib/sns.ts`	Wallet address resolved to `.sol` domain on connect; stored in on-chain memo at mint time

Onchain Data Sources (Tier 1)

DataBard has first-class adapters for onchain data — deep metadata extraction, no extra dependencies:

# The Graph — paste any subgraph URL
Source: The Graph (subgraph)
URL: https://api.thegraph.com/subgraphs/name/uniswap/uniswap-v3

# Dune Analytics — connect your query namespace
Source: Dune Analytics
API Key: <from dune.com/settings/api>
Namespace: uniswap

Bring Your Own Source (Tier 2 — Coral)

For sources we don't have a first-class adapter for yet, use Coral to connect anything via SQL:

# Coral — Join APIs, DBs, and local files with SQL
Source: Coral
Query: SELECT * FROM github.issues JOIN slack.messages ON ...

Coral supports 50+ sources (Salesforce, Jira, Postgres, Notion, Stripe, etc.) plus local CSV/JSON files. See docs/DATA_SOURCES_ARCHITECTURE.md for the full rationale behind tiered sources.

DataBard fetches query metadata via the Dune REST API, executes non-parameterized queries in parallel, and computes column statistics (min/max/avg, top values) from result data. This enables data-aware podcast narration — the episode discusses actual numbers and trends, not just schema structure.

Wallet Connect

Landing page → Connect Phantom/Solflare wallet
             → .sol domain resolved via SNS/Bonfida
             → Episodes minted on-chain under your .sol identity
             → Pay for Pro with Palm USD (29 PUSD/mo)
             → Health alerts registered per schema
             → Team history visible across wallets

What's Next

Palm USD Integration

DataBard accepts Palm USD (PUSD) as a native payment method for Pro subscriptions. Palm USD is a non-freezable, 1:1 USD-backed stablecoin on Solana — no admin key can seize, pause, or reverse a transaction.

Why Palm USD?

Non-freezable — no blacklist, no pause function, no clawback. Your subscription payment can never be censored.
1:1 USD backed — AED & SAR reserves held at regulated custodians, attested monthly (ISAE 3000).
Native on Solana — SPL token with sub-second finality. No bridges, no wrapped assets.
24/7 settlement — no weekends, no holidays, no cut-off times.

Payment Flow

User clicks "Pay with Palm USD"
  → Connect Solana wallet (Phantom, Solflare)
  → Server builds unsigned SPL token transfer (29 PUSD → DataBard treasury)
  → User signs in wallet
  → Transaction submitted to Solana
  → Server verifies on-chain confirmation
  → Pro access activated instantly

API Endpoints

Endpoint	Method	Description
`/api/checkout/palmusd`	POST	Builds unsigned PUSD transfer transaction. Body: `{ walletAddress }`
`/api/checkout/palmusd/verify`	POST	Verifies on-chain payment, activates Pro. Body: `{ walletAddress, txSignature }`

Configuration

# Solana network (mainnet-beta for production)
NEXT_PUBLIC_SOLANA_NETWORK=mainnet-beta
NEXT_PUBLIC_SOLANA_RPC_URL=https://api.mainnet-beta.solana.com

# Palm USD SPL token mint address (Solana mainnet)
NEXT_PUBLIC_PALM_USD_MINT=CZzgUBvxaMLwMhVSLgqJn3npmxoTo6nzMNQPAnwtHF3s

# Your treasury wallet that receives PUSD payments
PALM_USD_RECIPIENT=your_solana_wallet_address

Contract Details

Chain	Address	Type
Solana	`CZzgUBvxaMLwMhVSLgqJn3npmxoTo6nzMNQPAnwtHF3s`	SPL Token (6 decimals)

For more details, see docs/PALM_USD_INTEGRATION.md.

ElevenLabs Integration

DataBard uses ElevenLabs as its core audio engine — not just for TTS, but for the full production audio experience:

Voices

Host	Personality	ElevenLabs Voice	Voice ID
Alex	Enthusiastic data advocate	George	`JBFqnCBsd6RMkjVDRZzb`
Morgan	Skeptical quality auditor	Charlotte	`XB0fDUnXU5powFXDhCwa`

Both voices use context stitching (previous_text / next_text) for natural prosody across segment boundaries — the conversation flows like a real podcast, not a sequence of isolated TTS clips.

Sound Effects

Transitions between topics use ElevenLabs' Sound Effects API (textToSoundEffects.convert()). Each transition prompt is generated dynamically based on the topic shift:

// e.g. moving from quality failures to lineage section
await elevenlabs.textToSoundEffects.convert({
  text: "subtle data pipeline whoosh transition",
  duration_seconds: 1.5,
});

SFX are cached for 30 days to avoid redundant API calls.

Audio Pipeline

Script segments → ElevenLabs TTS (per segment, with context stitching)
                → ElevenLabs SFX (per transition)
                → Concatenated MP3 (streamed to client)
                → Cached 24hr

Setup

# ElevenLabs Starter plan ($5/month) required for API access
ELEVENLABS_API_KEY=sk_your_key_here

Tier	API Access	Best For
Free	❌ No API	Web UI only
Starter ($5/mo)	✅ Full API	Development & production
Creator ($22/mo)	✅ Full API + cloning	Custom voices

License

MIT

Name		Name	Last commit message	Last commit date
Latest commit History 106 Commits
.claude		.claude
.commandcode/taste		.commandcode/taste
docs		docs
public		public
scripts		scripts
src		src
.env.example		.env.example
.gitignore		.gitignore
.output.txt		.output.txt
KIRO.md		KIRO.md
LICENSE		LICENSE
README.md		README.md
ecosystem.config.cjs		ecosystem.config.cjs
eslint.config.mjs		eslint.config.mjs
next-env.d.ts		next-env.d.ts
next.config.mjs		next.config.mjs
package-lock.json		package-lock.json
package.json		package.json
postcss.config.js		postcss.config.js
tsconfig.json		tsconfig.json
tsconfig.tsbuildinfo		tsconfig.tsbuildinfo

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

🎙️ DataBard

The Problem

The Solution

How It Uses OpenMetadata

Architecture

Features

Quick Start

OpenMetadata sandbox mode (no user token required)

Tech Stack

Solana Integration

Solana Utilities

Onchain Data Sources (Tier 1)

Bring Your Own Source (Tier 2 — Coral)

Wallet Connect

What's Next

Palm USD Integration

Why Palm USD?

Payment Flow

API Endpoints

Configuration

Contract Details

ElevenLabs Integration

Voices

Sound Effects

Audio Pipeline

Setup

License

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

🎙️ DataBard

The Problem

The Solution

How It Uses OpenMetadata

Architecture

Features

Quick Start

OpenMetadata sandbox mode (no user token required)

Tech Stack

Solana Integration

Solana Utilities

Onchain Data Sources (Tier 1)

Bring Your Own Source (Tier 2 — Coral)

Wallet Connect

What's Next

Palm USD Integration

Why Palm USD?

Payment Flow

API Endpoints

Configuration

Contract Details

ElevenLabs Integration

Voices

Sound Effects

Audio Pipeline

Setup

License

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages