What does the dark pool trades MCP tool return?

It returns recent large dark-pool trade records with observed trade fields, available NBBO context, and clearly labeled derived fields. The output also includes the normalized query, market-date selection, retrieval time, pagination status, partial status, and warnings.

Can the dark pool MCP server search market-wide and by ticker?

Yes. Omit symbol to search market-wide, or provide one supported US ticker such as SPY, NVDA, or TSLA. A single call does not accept a list of symbols.

Does a dark-pool print reveal buy or sell direction?

No. A print does not prove buy or sell direction, institution identity, or trading intent. The tool can return a midpoint-side inference, but it is labeled as derived and must not be presented as observed fact.

How fresh are the dark-pool trade results?

The result states its freshness through marketDate, retrievedAt, and selection. The selection is either the requested date or the latest available session, so an agent should quote those fields instead of calling the data live.

Can AI predict stock prices with this MCP tool?

No. AI cannot predict stock prices with certainty, and this MCP tool does not promise future returns. It gives an agent bounded data or validation results that still require independent review, risk controls, and source checking.

Market analysis

Dark Pool MCP Server for Recent Large Prints

The Pineify dark pool MCP server lets an MCP-compatible AI agent find recent large dark-pool prints market-wide or for one US ticker. It returns observed trade and NBBO data with clearly labeled inferred side and size class. A print does not prove buy or sell direction, institution identity, or intent.

Connect Pineify MCP View all MCP tools

MCP tool name

find-dark-pool-trades

Example prompt

Use Pineify to find the 20 largest recent dark-pool trades market-wide by premium. Keep observed fields separate from inferred side.

Direct answer

What find-dark-pool-trades does

The find-dark-pool-trades tool answers bounded questions about recent large off-exchange prints in US stocks. Leave the symbol out for a market-wide result, or send one ticker such as SPY, NVDA, or TSLA. The agent can constrain the request by market date, premium, share size, market volume, inferred side, and derived size class.

Each result preserves the observed print fields separately from interpretation. Price, shares, premium, execution time, market center, cancellation status, and the available NBBO fields describe the record. Inferred side and size class are explicitly marked as derived and include a methodology version.

This separation matters. A dark-pool print confirms that a reported transaction met the query, but it does not disclose who traded, why they traded, whether the order opened or closed a position, or which participant initiated it.

Contract

What the agent sends and receives

Inputs

An optional symbol for one US ticker. The schema normalizes valid symbols to uppercase; omit it for a market-wide query.
An optional marketDate in YYYY-MM-DD format. Without it, the service selects the latest available market session.
Optional minimum and maximum values for premium in USD, trade size in shares, and market volume in shares.
An inferredSide filter of any, buy, sell, or neutral, plus a sizeClass filter of any, large-or-larger, block-or-larger, or mega.
A sort order of newest, premium-desc, or size-desc, and a limit from 1 to 50 with a default of 20.

Structured output

Up to 50 trade records with ID, execution time, symbol, price, share size, premium, market volume, market center, and cancellation status.
Available NBBO context for each record: bid, ask, bid quantity, and ask quantity. Individual fields can be null.
Derived inferred side and size class fields, each marked as derived and tied to a methodology version.
The normalized query, US market metadata, selected market date, retrieval timestamp, selection mode, and bounded page status.
A partial flag and bounded warnings when the service cannot return a complete result.

Capabilities

Where this MCP tool fits

Search market-wide or by ticker

Find recent large prints across the available US market dataset, or limit the call to one normalized ticker such as SPY.

Apply bounded filters

Constrain premium, share size, market volume, market date, inferred side, and size class, then sort by time, premium, or size.

Inspect trade and NBBO context

Review execution time, price, shares, premium, market center, cancellation status, bid, ask, and available quote quantities in structured output.

Keep inference labeled

Receive inferred side and size class as derived values with methodology versions, separate from the observed trade fields.

Agent workflow

A bounded call from question to review

1
Define a bounded question
Choose market-wide scope or one US ticker, then set only the date, premium, share-size, market-volume, side, and size filters needed for the review.
2
Call the read-only tool
Let the agent invoke find-dark-pool-trades with a limit of 50 or fewer rows and a sort order that matches the question.
3
Review observations and inferences
Check timestamps, NBBO fields, cancellation status, partial warnings, and methodology labels before discussing any possible interpretation.

Prompt examples

Questions an agent can route to this tool

"Use Pineify to find the 20 largest recent dark-pool trades market-wide by premium. Keep observed fields separate from inferred side."

"Use Pineify to review recent NVDA dark-pool prints with at least $1,000,000 in premium. Include execution time, shares, NBBO, and cancellation status."

"Use Pineify to find recent TSLA block-or-larger dark-pool prints. Explain why the derived side does not establish institutional intent."

Operator notes

How I review the result

When I review SPY, I sort by premium and compare each result against its execution time, share size, price, and available NBBO. I do not treat the largest print as proof of bullish or bearish direction.

For NVDA, I use one market date at a time and compare the execution price with the reported bid and ask. I keep the inferred-side label marked as derived and check its methodology version before citing it.

With TSLA, I narrow the result by premium or share size, then check cancellation status, retrievedAt, partial, and warnings. I record missing fields as missing instead of inventing a print or filling a gap.

Boundaries to keep in the prompt

A reported print does not reveal institution identity, order intent, position direction, or whether a trade opened or closed exposure. Midpoint-side inference is not guaranteed buy or sell direction.
The result is a bounded snapshot of the requested or latest available market date. It is not a live tape, a complete historical archive, or a reconstructed position.
Trade, NBBO, and market-center fields can be null. Partial results and warnings must remain visible in any summary.
The tool is read-only. It does not execute trades, modify a portfolio, predict prices, or provide investment advice.

Pineify MCP is an information and code-validation tool, not investment advice. It does not promise returns, place trades, or modify a portfolio. Review timestamps, source fields, code diagnostics, and risk assumptions before acting.

MCP tool family

Continue the workflow

Browse all 14 tools

Find Options Flow Alerts

Scan recent unusual US options-flow alerts with symbol and signal filters.

Get Market Tide

Review bounded market-wide options premium and its labeled sentiment inference.

Get Technical Analysis Snapshot

Read current technical-analysis fields for one supported ticker across selected timeframes.

Related Pineify tools

Market Insights

Institutional-grade market intelligence suite. Track real-time options flow, dark pool block trades, congressional stock disclosures, and net premium market sentiment in one unified dashboard.

Try it

FAQ