Slatemark Review: A Trade Journal Your AI Client Can Actually Read

Anyone who has tried to think through a trade with an AI client has run into the same wall: the model has no idea what you actually own, what your rules are, or what you were thinking when you opened the position. You end up copy-pasting positions into the chat, re-typing your framework rules from memory, and hoping the model doesn't invent numbers. Slatemark closes that gap. Instead of pasting context into a prompt, you connect your AI client to Slatemark once via the Model Context Protocol (MCP), and from then on the model can pull your journal entries, your rules, delayed quotes, and — on the Plus plan — your read-only brokerage positions. The model reasons over your data, not a generic summary of it.

Most AI-trading products are a variation on the same pitch: predictions, signals, buy/sell calls. Slatemark refuses that entire category. The tagline is deliberate: "No signals. Receipts." What that means in practice: Slatemark never tells you what to do. It hands your AI client the receipts — what you traded, when, why you said you'd trade it, what your rules were, what the market did around it. The AI does whatever reasoning happens; Slatemark just makes sure the inputs are real and traceable. Every tool result carries source and freshness metadata, so a Claude response can cite "SEC filing dated X" or "delayed quote fetched at Y" instead of confabulating. For traders who have watched AI products drift into black-box signal-selling, the restraint is refreshing. It's also the reason the compliance surface is small enough to actually be defensible.

Once Claude Desktop is wired to Slatemark, the AI stops guessing and starts referencing. A typical swing-trading day looks something like this: - **Pre-market:** "Give me a pre-trade brief on TSLA against my framework rules." Claude pulls the delayed quote, the technicals, and cross-checks against the position sizing and concentration caps you've saved. - **Mid-day:** "Review my open positions and flag anything that violates a rule I've set." Claude reads your read-only brokerage state (on Plus) and your framework, and answers with sources attached. - **After the close:** "Do a post-mortem on the NVDA trade I closed today against the thesis I wrote when I opened it." Claude reads the journal entry, the fills, and the outcome, and hands back a grounded write-up. Every response cites where the numbers came from and when they were fetched. You're not prompting an oracle — you're asking a very well-read assistant to reason over your actual book.

Slatemark ships with six pre-built workflows that map to the swing trader's week: - **Pre-trade brief** — delayed quotes and technical context for a ticker you're about to open - **Position review** — current exposure versus your concentration caps and framework rules *(Plus)* - **Earnings setup** — options snapshot heading into a print *(Plus)* - **Regime check** — cross-asset context including yield curve, dollar, and credit signals *(Plus)* - **Catalyst map** — dated primary-source research (SEC filings, FOMC schedule) - **Post-mortem** — closed-trade analysis against the thesis you actually wrote down Each workflow is a set of MCP tools the AI can call in sequence. You don't script anything — you ask Claude "walk me through my position review" and it pulls the relevant data. An optional senior-analyst skill on Claude adds a lightweight rule-checking layer on top of the raw tools.

The Scorecard is the payoff for keeping a disciplined journal. Once trades close, Slatemark computes win rate (with a confidence band), dollar expectancy, total P&L, and average holding period — filterable by any tag or rule you've applied to the trade. The filter dimension is where the product earns its keep. You can slice "only trades I tagged 'earnings-catalyst'" or "only setups where I violated my concentration rule" and see whether the pattern is actually working. Most trade journals show you a P&L number; Slatemark lets you interrogate why it looks the way it does.

The Free plan is more than a demo: manual journal entries, the Scorecard, one active AI-client connection, delayed market data, 30 requests per minute, and 2,500 successful tool calls per 30-day period. Free also includes pooled primary-source research — FRED, EIA, SEC EDGAR, CFTC, and Treasury — inside a shared fair-use budget. That's enough headroom for a full swing-trading workflow without ever touching a credit card. Plus at $4.99/month adds the piece that closes the manual-entry loop — periodic, read-only brokerage sync (accounts, balances, positions, and automatic import of fills into the journal). It also raises the rate limit to 60 req/min, opens additional AI-client connections, and adds short-interest history from FINRA and option-chain comparison across expirations. Billing is month-to-month via Stripe with a cancel-from-dashboard flow; if you downgrade, manual entries stay and broker-derived data is cleared. Premium bring-your-own-key sources for news and earnings are noted as not currently part of the public plans.