Build on the V5.4 Engine
One pick a day, 15 MCP tools, no auth. Point your AI agent, research notebook, or trading bot at the endpoint and go.
Connect Your Agent
The MCP API is free. No API key, no sign-up. Just connect and query.
https://gammarips-mcp-406581297632.us-central1.run.app/sseWant the pick pushed to WhatsApp and an AI agent in the group? Pro is $39/mo with a 7-day free trial →
Quick Start
Endpoint:https://gammarips-mcp-406581297632.us-central1.run.app/sse
Transport:SSE (Server-Sent Events)
Auth:None required
Primary tool:get_todays_pick
# Python (using fastmcp)
from fastmcp import Client
async with Client("https://gammarips-mcp-406581297632.us-central1.run.app/sse") as client:
pick = await client.call_tool("get_todays_pick", {})
print(pick)15 Available Tools
get_todays_pick is the primary entry point — it returns the single V5.4 pick for today (or null if the engine skipped). Everything else is supporting context.
get_todays_pick
PRIMARY
The single V5.4 pick for today's scan date. Returns ticker, direction, contract, entry/stop/target/exit, and the gate-pass evidence. Null when the engine skipped (no signal cleared the gates, or VIX backwardation).
Parameters
:No params required
list_todays_picks
Historical list of daily picks written by the notifier — one row per scan date.
Parameters
limit:integer (default 30)
get_freemium_preview
Redacted preview of today's pick for unauthenticated surfaces (ticker masked until 07:30 ET).
Parameters
:No params required
get_open_position
Composite payload about the engine's current state: pending pick, awaiting simulation, and most-recent closed trade. The batch simulator is not live; no fabricated unrealized P&L.
Parameters
:No params required
get_position_history
Closed trades from the forward-paper-trader ledger. Filters: date range, direction, outcome.
Parameters
start_date:YYYY-MM-DD (optional)
end_date:YYYY-MM-DD (optional)
direction:bullish | bearish (optional)
limit:integer (default 30)
get_overnight_signals
Raw overnight scanner rows across 5,230+ tickers.
Parameters
date:YYYY-MM-DD (optional)
direction:bull | bear (optional)
min_score:integer 1-10 (optional)
ticker:string (optional)
limit:integer (optional)
get_enriched_signals
Signals after the V5.4 enrichment filter (score ≥ 1, spread ≤ 10%, directional UOA > $500K). Includes thesis, technicals, recommended contract.
Parameters
date:YYYY-MM-DD (optional)
direction:bull | bear (optional)
ticker:string (optional)
limit:integer (optional)
get_signal_detail
Deep dive on a single ticker: full enriched payload including contract, thesis, technicals, news.
Parameters
ticker:string (required)
scan_date:YYYY-MM-DD (optional)
get_enriched_signal_schema
Returns the schema (column names + descriptions) of the enriched-signals table. Useful when an agent needs to know what fields to select.
Parameters
:No params required
get_signal_performance
Three-day forward returns per enriched signal. Separate universe from the paper-trader ledger — this is signal-level outcome tracking, not bracket trades.
Parameters
date:YYYY-MM-DD (optional)
ticker:string (optional)
direction:bull | bear (optional)
outcome:win | loss (optional)
limit:integer (optional)
get_win_rate_summary
Aggregate signal-level win rate across a rolling window. Note: this is the signal universe (~30/day), not the V5.4 paper-trader universe (1/day).
Parameters
days:integer (default 30)
get_daily_report
Full daily market report (AI-authored editorial synthesis) in markdown.
Parameters
date:YYYY-MM-DD (defaults to latest)
get_report_list
Paginated list of available daily reports.
Parameters
limit:integer (default 10)
get_available_dates
Returns which scan dates have data.
Parameters
:No params required
web_search
Web search for real-time fact verification. Requires GOOGLE_CSE_ID at deploy time.
Parameters
query:string (required)
num_results:integer (optional)
Pricing for Developers
FREE
MCP API
$0
- ✓ All 15 tools unlocked
- ✓ No auth, no API key
- ✓ SSE transport
- ✓ Rate-limited fairly; reach out for heavy use
- ✓ Same data as the webapp, same second
Ready to Build?
Point your agent at the endpoint and start querying. That's it.