Slatemark.
Get started

No signals. Receipts.

A read-only record for the AI client you already use.

Slatemark keeps the record behind every trade: capture why it exists, compare the facts with your own recorded rules, record the outcome, and review the pattern across your closed trades. Your AI client creates the response. Slatemark hosts no model and cannot place orders.

Trade journal. Read-only MCP. Catalyst calendar and Weekly Slate are not yet available.

Built first for swing traders managing multi-day to multi-week trades across brokerage accounts.

Bring your own AI client. Works with Claude Desktop, Claude Code, Cursor, and more. The optional senior-analyst skill is a file you choose to install in that client.

Read-only · not investment advice · no performance claim. Read the terms.

One factual comparison

Hand-authored fictional example. No live tools or model generated this transcript.

User: “Compare my NVDA exposure with the 25% concentration cap I recorded.”

Your AI alone

“I do not have access to your brokerage position or the cap you recorded. Share those values if you want a comparison.”

A clear limitation: the client has no account or journal context to compare.

Example response from your AI client

“In this fictional record, NVDA is 23.4%periodically synced Account Data of Net Liquidation Value: the total cash-equivalent value of your account if every open position were closed at the current mark.. Your recorded cap is 25%framework rule. The difference is 1.6 percentage points below that cap.

The delayed market-data quote is $1,043.20approximately 15 minutes delayed. This reports the comparison only; it does not prescribe an action.”

Slatemark supplies source, fetch-time, and freshness metadata where available. The user's AI client decides how to present those fields.

How it works

1
Capture the plan

Log why the trade exists and what would invalidate it, in the dashboard trade journal or through your AI client. You still add the thesis, rules, tags, and notes to your journal.

2
Compare the facts with your rules

Your AI client calls Slatemark's read-only tools for delayed market data, primary-source research, your framework rules, and periodically synced Account Data, then creates the response. Slatemark does not host the model or own its reply.

3
Record the outcome

Close the entry with what actually happened, without rewriting the original intent. Manual and broker-derived facts stay distinguishable.

4
Review the pattern

The Strategy Scorecard reads your own closed trades by the tags you chose: win rate, expectancy, and P&L. Descriptive of the past, never a forecast.

Six moves to start with

These six facts-only workflows ship as slash commands in the Claude plugin, and the same prompts work in any connected AI client. The external AI client creates the response from Slatemark tool results.

Pre-trade brief

Delayed market context. Report a delayed quote, moving averages, RSI, ATR, and their relationship to a level the user supplied.

Position review

Recorded position context. Compare exposure with a cap the user recorded and quote the user's rule without prescribing an action.

Earnings setup

Direction-neutral option snapshot. Report a delayed ATM straddle and approximate implied range where listed, with no Greeks and no structure recommendation.

Regime check

Cross-asset context. Report the curve, dollar, credit, financial conditions, and volatility as observed data.

Catalyst map

Dated source research. Collect available earnings, guidance, and macro dates without turning the timeline into an instruction.

Post-mortem

Past closed-trade review. Separate recorded outcomes from the user's notes, then preserve the user's own lesson and tags.

What you get

Trade journal

Keep the user's intent beside the recorded outcome. Add a thesis, rules, tags, and notes. Available brokerage activity can be reconciled periodically, but the user's reasoning remains a deliberate input.

Read-only MCP Plus

Give the AI client you already use a data pipe. It can request delayed market data, primary-source research, journal entries, framework rules, and authorized Account Data. Slatemark supplies tool results; the client writes the reply.

Clear data boundaries

Market data and Account Data are separate. Market data is approximately 15 minutes delayed. Brokerage connections provide periodically synced balances, positions, orders, and activity, not market data. Nothing writes back to a brokerage.

Strategy Scorecard, on your past trades

Review closed-trade outcomes by the tags and rules you recorded: win rate with a confidence band, dollar expectancy, total P&L, and holding period. Available brokerage activity is reconciled periodically; you add the tags, rules, and notes that make the comparison meaningful. This is a description of past activity, not a forecast or recommendation. See how it works.

No signals. Receipts.

Slatemark supplies read-only data and keeps the record you create. It does not tell you what to trade, and it does not generate the response you read in your AI client.

  • Read-only, by construction. No order entry. The brokerage link runs through an aggregator on read-only permissions. Slatemark cannot place, modify, or cancel an order, move money, or change your account.
  • No signals, no trade ideas. Nothing on any surface says what to trade or when.
  • No real-time market data. Market data is approximately 15 minutes delayed and labeled with its freshness. Brokerage connections provide Account Data, not market data.
  • Your AI client owns the reply. Your external AI client creates the response from read-only tool results it chooses to use. Slatemark hosts no model. Starting on Plus, the client reaches a read-only MCP connection (Model Context Protocol, the open standard AI clients use to call data tools) to delayed market data, primary-source research, framework rules, and brokerage Account Data. The optional senior-analyst skill is installed by you in that client.
  • Provenance is supplied, not guaranteed as prose. Tool results include source, fetch time, freshness, and related metadata where available. Slatemark cannot guarantee how another AI client cites or presents it.
  • Disconnecting has a precise data boundary. Disconnecting a brokerage connection deletes broker-derived data. Journal entries you wrote manually remain.
  • Plans. The dashboard trade journal is available on Free. MCP access for an external AI client starts on Plus.

Ready to connect a read-only record?

Start on the Free plan with the dashboard trade journal, no card and no trial timer. MCP access starts on Plus.