Examples - Allium Documentation Hub

These examples follow the questions customers ask the Assistant most often. Each one shows how the Assistant chains its tools together, from the first question to the final result.

Investigate

Identify a wallet and trace its funds

You have an address from an alert, a counterparty list, or a block explorer, and you need to know who it is and where its money moves.

Ask who owns the wallet

Who owns this wallet? 0x28C6c06298d514Db089934071355E5743bf21d60
Show where its USDT came from and where it went over the last 30 days.

Look up the address labels

The Assistant looks up the address and returns its name, category, and project. This address is labelled as a Binance hot wallet, so the Assistant treats it as an exchange wallet in the rest of the analysis.

Trace inflows and outflows

The Assistant finds the token transfer tables for Ethereum, then writes SQL that sums USDT transfers into and out of the wallet by counterparty over the last 30 days.

Label the counterparties

It looks up labels for the largest counterparties, so each one shows as an exchange, a bridge, a DEX, or an unlabelled wallet.

Get the wallet profile

The Assistant returns two tables, top sources and top destinations, each with the USD amount, transfer count, and label. It then summarizes the main flows and asks if you want to follow the largest unlabelled counterparty one hop further.

Investigate unusual activity on a given day

A chart shows a spike, and you need to know what caused it before someone asks.

Describe what you saw

DEX volume on Base spiked on January 20, 2026. What happened? Find the pools and the wallets behind it.

Asking Allium Assistant what caused a DEX volume spike on Base

Confirm the spike

The Assistant queries daily DEX volume on Base for the days around that date and confirms the spike in text: $1.4B on January 14, rising to $3.57B on January 20, then falling to $633M by January 24.

Break the day down

It splits that day’s volume by pool and pair, then by the wallets that sent the trades (transaction_from_address), to find where the extra volume came from.

Label the wallets

The Assistant looks up labels for the top wallets, so you can tell a bot from a market maker or an aggregator. In this recording, the second wallet in the list is labeled as an MEV sandwich trader.

Get the explanation

The Assistant explains the cause in a few lines and shows the tables that support it. Here, one wallet wash-traded a single pair, ROLL/USDC, in one Uniswap V4 pool. That wallet made $2.11B, 59% of the day’s volume, and the day is about $1.46B without it. The Assistant then offers to save the queries as Explorer queries and build a dashboard, or to trace where the wallet’s USDC came from and went.

Debug queries

Fix a query that fails

You wrote a query in Explorer, or copied one from a colleague, and it returns an error.

Paste the error and the query

My query failed with the following error:
SQL compilation error: invalid identifier 'USD_VOLUME'

SELECT DATE(block_timestamp) AS day, chain, SUM(usd_volume) AS volume
FROM crosschain.dex.trades
WHERE block_timestamp >= CURRENT_DATE - 7
GROUP BY 1, 2
ORDER BY 1

Asking Allium Assistant to fix a failing query

Check the table schema

The Assistant searches the schema for crosschain.dex.trades and finds that the USD value of a trade is in usd_amount. There is no usd_volume column.

Rewrite the query

SELECT
    DATE(block_timestamp) AS day,
    chain,
    SUM(usd_amount) AS volume
FROM crosschain.dex.trades
WHERE block_timestamp >= CURRENT_DATE - 7
GROUP BY 1, 2
ORDER BY 1

Run it or save it

The Assistant explains the change in one line. It then offers to run the query, or to save it as a reusable Explorer query and chart daily DEX volume by chain.

You follow the tokenized equities market and want a dashboard your team can open every morning.

Ask for the trend and the dashboard

Show daily trading volume of tokenized stocks over the last 90 days, split into spot and perps, and build a dashboard I can share.

Asking Allium Assistant for a shareable tokenized stocks dashboard

Check the metrics catalog

The Assistant checks Allium’s curated metrics and Terminal dashboards first. The RWA catalog has spot and perps volume metrics that split by asset class, and tokenized stocks are the equities asset class.

Find the tables and save a query

The Assistant searches the schemas and finds rwa.core.spot_trades and rwa.core.perpetual_trades. It validates a query that unions daily spot and perps volume for equities over the last 90 days, runs it, and saves it to Explorer. The saved query returns 182 rows, one per day for each of the two series.

Chart it in the conversation

The Assistant draws a stacked area chart of daily volume, with one band for spot and one for perps.

Build the dashboard

The Assistant builds a dashboard titled “Tokenized Stocks Trading Volume - Spot vs Perps” with:

A markdown header that explains the scope and the two source tables
Three big-number cards for total, spot, and perps volume over 90 days: $476.27B, $42.45B, and $433.82B
A full-width stacked area chart of daily spot and perps volume, with a daily granularity picker

Share the link

The Assistant makes the dashboard public and returns the link, so anyone with the link can open it without logging in. It adds the takeaways: perps are routinely 5 to 15 times spot volume, perps fall under $1B on weekends and reach $4B to $10B or more on weekdays, and spot has trended up through September. It then offers a spot share view, a longer window, or a post for X.

Build on previous context

A multi-turn conversation in which each follow-up builds on the last result, without repeating the context.

Turn 1: Ask the first question

Build me monthly Hyperliquid perp volume by category (crypto, equities, commodities) for this year.

The Assistant asks two short questions first: how to classify markets into categories, and what “this year” means. Once you answer, it finds the Hyperliquid perpetual trade tables, groups markets by category, and returns a stacked bar chart of monthly volume.

Building from one chart to a shared dashboard over five turns

Turn 2: Narrow the scope

Change it to the S&P 500 perp, daily, since launch.

The Assistant asks which S&P 500 perp you mean. Once you pick one, it edits the previous query: it filters to that market, changes the grain to daily, and starts at the market’s launch date. The edit replaces the turn 1 query.

Turn 3: Add a second metric

Add the number of unique traders and plot it on a separate axis.

The Assistant adds a distinct trader count to the query and redraws the chart with unique traders as a line on a right-hand axis.

Turn 4: Visualize

Create a dashboard with both charts.

The Assistant recreates the category chart as its own query, because turn 2 replaced the original. It then builds a dashboard with the category chart and the S&P 500 chart from turn 3.

Turn 5: Share

Make it public and share on X.

The Assistant makes the dashboard public, writes a post with the key takeaways and the dashboard link, and returns an X link with the post pre-filled. You review and post it.

​Investigate

​Identify a wallet and trace its funds

​Investigate unusual activity on a given day

​Debug queries

​Fix a query that fails

​Analyze and share

​Track tokenized stocks and share a dashboard

​Build on previous context