fahali-mcp
Fahali MCP is an uncertainty-aware market-intelligence server providing provenance-backed financial analysis across ~900 crypto and US equity/ETF instruments. Every response includes a confidence score and reasoning chain; uncalibrated engines report 'calibrating' rather than inventing numbers.
Market Verdicts (
fahali_get_market_verdict) — Structured bullish/bearish/neutral direction, confidence score, reasoning chain, expected move %, and forecast horizon for one or more symbols. Optionally enriched with contagion risk data. Available on Free (limited, 50 calls/day) and Professional tiers.Portfolio Risk (
fahali_get_portfolio_risk) — Overall risk score, Value-at-Risk (VaR), expected drawdown, and per-position breakdown. Requires connected exchange accounts and Elite tier.Contagion Maps (
fahali_get_contagion_map) — Cross-asset tail-dependence map showing nodes (symbols with risk scores), pairwise correlation edges, and clusters of correlated assets. Requires Elite tier.Capital Flow Intelligence (
fahali_get_capital_flow) — Net USD flow, institutional vs. retail breakdown, whale activity, and overall flow trend for a specific symbol. Requires Professional tier.
Access: Connect remotely via SSE with API key authentication or run locally as a stdio MCP server. All outputs are informational only and not financial advice.
Provides CrewAI adapter to integrate market-risk intelligence tools into CrewAI multi-agent systems.
Provides LangChain adapter to integrate market-risk intelligence tools into LangChain agent workflows.
Allows ChatGPT to access market-risk intelligence tools using the OpenAI tools format, providing structured market judgment and reasoning.
Provides adapters for Vercel AI SDK to integrate market-risk intelligence tools into AI applications.
Fahali MCP Server
Market-risk intelligence your AI agent can call. Fahali gives agents structured, read-only judgment about markets — direction, confidence, reasoning, expected move, contagion, crash precursors — instead of raw prices they have to interpret blindly.
Informational only — not financial advice. Fahali observes and explains risk. It never tells you or your agent what to buy or sell. No order routing, no path to capital.
What makes it different
Most market APIs hand an LLM data exhaust and hope it reasons well. Fahali hands it a verdict with a receipt:
18 detection engines scan ~840 instruments per cycle across a 2,000+ instrument universe — crypto 24/7 and US equities/ETFs during market hours. Anomaly, capital flow, market regime, dark-pool proxy, whale activity, funding stress, liquidation cascade, tail-dependence/contagion, and 72-hour crash-precursor forecasting.
An adaptive ensemble re-weights those 18 engines every 15 minutes based on which ones have been right lately, per market regime.
Every served claim is judged against realized outcomes and gets a permanent, public replay URL — hits and misses. The full grading methodology is published at fahaliai.com/methodology. Engines without enough resolved outcomes report
calibrating— never an invented number.
That last point is the whole thesis: an agent (or the human behind it) can audit any Fahali claim against what the market actually did. Example receipts, misses included, live at app.fahaliai.com.
Related MCP server: carbon-cashmere-mcp
Connect
Remote MCP (recommended — nothing to install):
Streamable HTTP: https://mcp.fahaliai.com/mcp
SSE (legacy): https://mcp.fahaliai.com/sseAuth: OAuth 2.1 ("Connect with Fahali") or a static key — Authorization: Bearer sk_live_.... Grab a free developer key (50 calls/day) at app.fahaliai.com/developer.
Claude / ChatGPT / Cursor: add the remote MCP URL above as a connector. Listed in the official MCP registry as com.fahaliai/fahali.
SDKs for building agents directly:
npm install fahali # TypeScript — OpenAI tools format + Vercel AI SDK / LangChain adapters
pip install fahali # Python — LangChain / CrewAI / LlamaIndex adaptersTools (25)
Market-intelligence tools are strictly read-only. A few optional agent-memory and custom-alert tools write only to your own agent workspace, never to markets.
Tool | Returns |
| Per-symbol verdict: direction, confidence, reasoning chain, expected move, horizon |
| Probabilistic 72h forecast: crash/neutral/pump probabilities, expected return, uncertainty cone |
| Portfolio risk: score, VaR, drawdown, per-position breakdown |
| Flash-crash precursor signals |
| Large-order / whale flow |
| Off-tape absorption (proxy) |
| Cross-asset tail-dependence / correlated clusters |
| Net flow, institutional vs retail, per symbol |
| Current regime read (HMM-based) |
| The judged record with per-horizon base rates |
| Stress-test a portfolio against a scenario |
… | 14 more — market snapshot, sentiment, correlation matrix, case studies, engine status, briefing, and agent-workspace tools |
Full tool schemas: mcp.fahaliai.com.
Pricing
Free developer key (50 calls/day) for exploration. Agent lanes from $49/mo (10k calls) → $199 (100k) → $999 (1M). Human app tiers from $19/mo at fahaliai.com.
Links
App: app.fahaliai.com
MCP docs / health: mcp.fahaliai.com
Methodology (how signals are graded): fahaliai.com/methodology
Developer / API: fahaliai.com/developer
Privacy: app.fahaliai.com/privacy
Built by Future Legends Inc. Observation, not advice.
Maintenance
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Future-Legends-Inc/fahali-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server