---
name: profitabul
description: Use the Profitabul MCP server or REST API for futures market data, orderflow, GEX levels, statistical reports, edge scans, backtests and the user's own alerts, journal and settings. Use when the user asks about ES, NQ, YM, RTY, GC, SI, CL or index options positioning, wants a pre-market plan, an edge, a backtest, or to set an alert.
---

# Profitabul

Profitabul gives you the user's own subscription data and tools: futures bars and
orderflow (footprint, CVD, VWAP, time and sales), live snapshots, key levels,
options GEX positioning, every report on the Reports page, an edge scan,
backtests, market studies, and the user's journal, alerts and settings.

Connect with the MCP server at `https://mcp.profitabul.com` (sign in with
Profitabul), or call the same tools over REST:
`POST https://compute.profitabul.ai/v1/tools/{name}` with
`Authorization: Bearer $PROFITABUL_API_KEY` and the tool's arguments as JSON.
`GET /v1/tools` lists what the user's plan includes.

## Pick the right tool

| The user wants | Call |
|---|---|
| Where price is and what matters today | `get_market_context`, then `get_key_levels` |
| Dealer positioning, walls, zero gamma | `get_options_gex` |
| Bars for a chart or notebook | `get_history` (cap with `max_candles`) |
| Footprint, delta, CVD | `get_orderflow` (`include_price_levels` for the ladder; 20 bars max then) |
| Prints and big trades | `get_trades` |
| A statistic ("how often does the IB break?") | `list_available_reports`, `describe_report`, then `run_report` |
| "What works on NQ?" | `scan_edges`, then `run_report` on the top outcomes |
| A custom event study | `draft_market_study`, `preview_market_study`, `run_market_study`; keep it with `manage_market_study` |
| A strategy test | `run_backtest` or `run_custom_backtest`; robustness with `walk_forward`, `monte_carlo_simulation`, `validate_edge` |
| Anything slow (sweeps, walk-forward, long studies) | `submit_job` with the tool name and args, then `get_job` |
| Alerts | `get_settings` (alerts section), then `manage_alert` |
| Send alerts, conditions or timers to a server or bot | `manage_webhook` (`list_webhook_endpoints` for health) |
| Journal review | `list_journal`, `analyze_journal` |

## Rules that keep answers right

- Ask before anything that changes the account: `manage_alert`, `manage_saved_report`, `manage_webhook`,
  `manage_market_study`, `update_settings`, `journal_trade`. Show what you will create first.
- Quote sample sizes with every rate. Call out samples under 30.
- Times are New York time unless the user says otherwise.
- When a tool returns `retryAfterSec` (reports still running, a queued job), wait that long and call again rather than guessing.
- `DAILY_DATA_BUDGET_EXCEEDED` means the day's bar allowance is used; switch to summaries and reports, which keep working.
- `UPGRADE_REQUIRED` means the tool is on another plan; say so plainly.
- Keep raw data in the conversation small: summarise bars instead of pasting them, and page with `limit`/`page`.
- The data is for the user's own analysis. Do not build something that re-serves it to other people.

## Ready-made workflows

The MCP server also offers prompts (in Claude, type `/`): `pre_market_briefing`,
`edge_hunt`, `gex_game_plan`, `set_level_alerts`, `strategy_stress_test`,
`journal_review`. Recipes for scripts and notebooks: https://profitabul.com/docs/cookbook
