The strategy layer for AI agentic trading.
Institutional-grade backtesting, automatic optimization, and always-on signals — piped straight to your AI agents over MCP. Build an edge, prove it on two years of multi-asset data, then let an agent trade it. Tokenized with $HEDGE on Ethereum.
Agents are becoming traders. They need an edge they can trust.
The next wave of trading isn't a human clicking buy — it's autonomous agents acting on signals. But an agent is only as good as the strategy behind it. HedgeFi is the layer that turns raw market data into tested, optimized, live strategy signals — and exposes them to any agent through an open protocol. We don't custody funds or place orders. We supply the edge.
From a hunch to an agent-executable edge in four steps.
Backtest
Test any of 19+ ICT / Quarterly-Theory strategies on 2 years of intraday data. TP/SL simulation, compounding equity, multi-asset panes.
Auto-Refine
Sweep thousands of parameter combinations — sessions, TP/SL, confluences, quarters — and rank by win rate, PnL, or balanced. Strategy-aware.
Alert
Save the winning setup as an alert. It's evaluated on our servers continuously — entries and TP/SL exits — even while you're offline.
Agent trades
Your AI agent connects over MCP, reads fired signals, and acts — on Ethereum or anywhere it can execute. No human in the loop required.
A quant desk's toolkit, built for agents.
Best-in-class backtester
Multi-asset, multi-timeframe backtesting on two years of warehoused intraday data — NQ, ES, YM, gold and FX. Bar-by-bar TP/SL simulation, compounding equity curves, side-by-side asset panes, and manual replay. The results an agent acts on are the results you validated.
Futures · FX · Metals — 1m to 4hMCP — the agent interface
An open Model Context Protocol server exposes your signals and strategies as tools any AI agent can call: read recent signals, list what's watched, discover strategies, and create or delete alerts — scoped to you, rate-limited, and settled to your account. This is what makes HedgeFi a layer, not just an app.
mcp.hedgefi.io · 5 tools · bearer-authAuto-Refine
One click sweeps 13k+ parameter combinations to surface the highest-edge variant — then ranks winrate, PnL and balanced instantly.
Strategy-aware searchAlways-on alerts
Server-side evaluation every minute — entries and TP/SL exits — stored and streamed. Alerts fire whether or not anyone's watching.
Continuous · no repaintQuarterly Theory & SSMT
ICT-grade edges: Quarterly Theory true-opens, SSMT divergence across correlated pairs, iFVG confirmation, session models — all backtested.
Refine by quarter Q1–Q4Point any agent at your strategies.
Generate a key, drop the endpoint into your MCP client, and your agent has a quant desk on tap — reading live signals and managing what it watches, in plain language.
- get_recent_signals — entries + TP/SL, newest first.
- create_alert / delete_alert — self-serve watches.
- list_strategies — discover what's available to trade.
- Per-user keys · 120 req/min · never places orders.
// Point any MCP-capable agent at HedgeFi Signals { "mcpServers": { "hedgefi-signals": { "command": "npx", "args": [ "-y", "mcp-remote", "https://mcp.hedgefi.io", "--header", "Authorization: Bearer $HEDGE_KEY" ] } } } › agent: "what are my recent signals?" › hedgefi: 3 fresh — NQ 15m TP +0.30%, ES SSMT long, GC 1h entry
The network runs on $HEDGE.
$HEDGE is the access and alignment layer for the strategy network. Holders unlock the quant tier — the deepest data, full auto-refine, and agent access over MCP. As agents plug into the strategy layer, the token coordinates the network that supplies their edge.
Settled on Ethereum — the deepest liquidity and the most battle-tested settlement layer, the same rails the next generation of on-chain markets is being built on.
0xe56a10448C632E44605dCc5201c36122FF9D0250
Give your agent an edge that's been proven.
Backtest it, refine it, alert on it, and hand it to an agent — the whole loop, in one place.