Skip to main content
The Allium AI Assistant is an agentic blockchain data analyst built into app.allium.so. It goes beyond simple SQL generation - it can search schemas, write and execute queries, create dashboards, share results, look up wallets, generate PDF reports, and even deploy live web apps, all through natural language conversation.
AI Assistant chat interface showing a multi-step analytical workflow

​
Capabilities

Natural language to SQL

Schema discovery, SQL generation, validation, and execution with confirmation flow

Dashboards & visualizations

Create interactive dashboards with charts, then share them publicly

Sharing & export

Public links, share to X/LinkedIn, PDF reports

Real-time data

Token prices, wallet balances, and transaction history - no SQL needed

Organization memory

Persistent context shared across your entire org

Web search

Current events, market context, and protocol research

Data pipelines (Beam)

Create real-time data pipelines with natural language

Live web apps (Vercel)

Deploy interactive Next.js apps from conversation

​
Getting started

1

Open the Assistant

Click the Assistant icon in the sidebar of the Allium App.
AI Assistant sidebar panel in Allium App
2

Ask a question

Type your question in plain English. The assistant plans and executes the steps needed to answer it.Example questions:
  • “Show me DEX volume by chain for the last 30 days”
  • “What are the top 10 tokens by trading volume on Ethereum?”
  • “Create a dashboard showing daily active wallets on Solana”
  • “What’s the current price of ETH?”
  • “Look up this wallet: 0xabc…”
3

Iterate and refine

Continue the conversation to drill deeper. The assistant remembers your full session context, so you can say things like “break that down by DEX” or “add a chart for that” without repeating yourself.

​
Example workflow

Question:
Show me DEX trading volume by blockchain over the last 90 days and create a dashboard for it.
What the assistant does:
  1. Searches schemas to find the right tables (crosschain.dex.trades)
  2. Writes and validates SQL:
SELECT
    DATE(block_timestamp) AS date,
    chain,
    SUM(usd_amount) AS usd_volume,
    COUNT(DISTINCT transaction_from_address) AS traders
FROM
    crosschain.dex.trades
WHERE
    block_timestamp >= CURRENT_TIMESTAMP - INTERVAL '90 days'
GROUP BY
    date, chain
ORDER BY
    date DESC;
  1. Executes the query and returns results
  2. Creates a dashboard with charts visualizing volume and trader counts by chain
  3. Shares a link so you can access the dashboard anytime
See Examples for more detailed end-to-end walkthroughs.

​
Tips

Be specific

The more precise your question, the better the assistant can find the right schemas and write accurate queries.

Let it work

The assistant handles multi-step workflows automatically. Ask for the end result and let it figure out the steps.

Iterate naturally

Build on previous results in the same conversation - refine queries, add filters, or request visualizations.

Review results

Always review generated queries and dashboards to ensure they match your requirements.

​
Need help?

For specific queries or deeper insights beyond the AI Assistant, reach out to support@allium.so. We’re happy to set up a Slack or Telegram channel for easy communication.