Use Case

LLM Trading Agents

Give Claude, ChatGPT, or your custom LLM real-time market context. Structured indicator data + natural-language summaries = LLM agents that reason about markets, not hallucinate about them.

The Challenge

LLMs are excellent at reasoning, but they have no access to real-time market data. Asking ChatGPT "should I buy EURUSD?" gets a generic disclaimer, not a data-driven answer. To build a useful trading agent, you need to ground the LLM with current, factual market data — indicator values, price action, volatility metrics, and market sentiment. The data needs to be structured enough for the LLM to interpret, and you need a natural-language summary it can synthesize into advice.

How TickAtlas Solves It

LLM-Ready /v1/summary

Natural-language market analysis with bias, confidence, key levels, and supporting evidence. Drop straight into your LLM prompt.

Structured Indicator Data

42 indicators as clean JSON. LLMs can parse "RSI: 72.3, MACD histogram: -0.0015" better than raw OHLCV candles.

Function Calling / Tool Use

Define TickAtlas endpoints as LLM tools. The agent decides when to check RSI, scan for setups, or get a summary.

Multi-Symbol Context

/v1/multi fetches indicators across multiple pairs in one call. Give your LLM a complete market picture in a single prompt injection.

Key Endpoints

GET /v1/summary?symbol=EURUSD

Natural-language market analysis — the primary LLM context source

GET /v1/multi?symbols=EURUSD,GBPUSD,XAUUSD&indicators=RSI_14,MACD_hist,ADX

Structured multi-symbol data for broad market context

GET /v1/screener

LLM tool: "find symbols matching these conditions"

GET /v1/indicator

LLM tool: "check this specific indicator for a symbol"

LLM Agent Architecture

architecture

┌──────────────────────────────────────────────────────────────┐
│                    LLM Trading Agent                         │
│                                                              │
│  ┌───────────────┐    ┌──────────────┐    ┌──────────────┐   │
│  │ User Prompt   │    │ LLM (Claude/ │    │ Tool Results │   │
│  │ "Analyze gold │───▶│  ChatGPT)    │◀───│ from API     │   │
│  │  for me"      │    │              │    └──────────────┘   │
│  └───────────────┘    │  Reasoning:  │           ▲           │
│                       │  1. Get summary│          │           │
│                       │  2. Check RSI  │          │           │
│                       │  3. Scan market │         │           │
│                       │  4. Recommend  │          │           │
│                       └──────┬───────┘           │           │
│                              │ Tool calls         │           │
│                              ▼                    │           │
│                       ┌──────────────┐            │           │
│                       │ Tool Router  │────────────┘           │
│                       │              │                        │
│                       │ get_summary()│──▶ /v1/summary         │
│                       │ get_rsi()    │──▶ /v1/indicator       │
│                       │ scan_market()│──▶ /v1/screener        │
│                       └──────────────┘                        │
└──────────────────────────────────────────────────────────────┘
                               │
                               ▼
                    TickAtlas API (https://tickatlas.com/v1)

Code Example: Claude Tool Use

python

import anthropic
import requests

CLAW_KEY = "your_tickatlas_key"
BASE = "https://tickatlas.com/v1"
headers = {"X-API-Key": CLAW_KEY}

# Define TickAtlas as a tool for Claude
tools = [
    {
        "name": "get_market_summary",
        "description": "Get a rule-based market summary with bias and confidence for a forex/crypto symbol",
        "input_schema": {
            "type": "object",
            "properties": {
                "symbol": {"type": "string", "description": "e.g. EURUSD, XAUUSD, BTCUSD"}
            },
            "required": ["symbol"]
        }
    },
    {
        "name": "get_indicator",
        "description": "Get a technical indicator value for a symbol and timeframe",
        "input_schema": {
            "type": "object",
            "properties": {
                "symbol": {"type": "string"},
                "indicator": {"type": "string", "description": "e.g. RSI_14, MACD_hist, BB_upper"},
                "timeframe": {"type": "string", "description": "e.g. M15, H1, H4, D1"}
            },
            "required": ["symbol", "indicator", "timeframe"]
        }
    }
]

def handle_tool_call(name, inputs):
    if name == "get_market_summary":
        return requests.get(f"{BASE}/summary", headers=headers,
            params={"symbol": inputs["symbol"]}).json()
    elif name == "get_indicator":
        return requests.get(f"{BASE}/indicator", headers=headers,
            params=inputs).json()

# Ask Claude to analyze a symbol using real data
client = anthropic.Anthropic()
response = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    tools=tools,
    messages=[{"role": "user", "content": "Analyze XAUUSD for a potential swing trade entry"}]
)

# Claude will call get_market_summary and get_indicator tools,
# then synthesize a data-grounded trading analysis

Recommended Plan

Pro Plan

$79/mo

LLM agents make multiple tool calls per user query. Pro plan gives 100,000 requests/day for the /v1/summary and indicator calls that are the key context source for LLM reasoning.

✓ /v1/summary for LLM-ready market analysis
✓ 100,000 requests/day for multi-tool agent workflows
✓ 10 API keys for multiple agent instances

View all plans →

Ground Your LLM in Real Data

Start free with $2.50 of API credit. Build an AI trading agent that reasons with facts, not hallucinations.

Start free SDKs & Libraries