How to Call Quan 3.4, Quan 3.4 L, and Deep Research APIs

Quick Start

All Quan API calls use the same endpoint: https://stockup.cc/v1/query. Create a key in the StockUp Developer Console, pass it through the x-api-key header, and send OpenAI-style chat messages in JSON. New verified developers receive 100,000 free Quan 3.4 L tokens without a card or prepaid balance. Quan 3.0 keeps a separate 100,000-token allowance.

curl -X POST https://stockup.cc/v1/query \
  -H "Content-Type: application/json" \
  -H "x-api-key: STOCKUP_API_KEY" \
  -d '{
    "model": "quan-3.4",
    "messages": [
      {
        "role": "user",
        "content": "Compare AAPL and MSFT valuation today. Separate confirmed data from interpretation."
      }
    ],
    "stream": false,
    "webSearch": true,
    "maxOutputTokens": 65536,
    "temperature": 0.2
  }'

Which Model Should You Call?

Calling Quan 3.4 L

Quan 3.4 L includes 100,000 account-level tokens for verified developers. It supports the same request shape, streaming, verified StockUp context, and opt-in live grounding as Quan 3.4. A prepaid wallet is needed only for token overage, searches after the two included per grounded free request, or continued use after the allowance.

curl -X POST https://stockup.cc/v1/query \
  -H "Content-Type: application/json" \
  -H "x-api-key: STOCKUP_API_KEY" \
  -d '{
    "model": "quan-3.4-l",
    "messages": [
      { "role": "user", "content": "Summarize today\'s U.S. market move. Separate verified facts from inference." }
    ],
    "stream": false,
    "webSearch": true
  }'

Light Market Insights

For easy ticker questions like How is AAPL performing today?, Quan 3.0 and Quan 3.4 now use a Light Market Intent layer. The default response is still plain text, but Quan can add a concise source-grounded read: price, session move, move character, and a useful next thing to watch.

If your app wants renderable chips or cards, set includeInsightMetadata: true. For stream: false, the API may include a top-level quanLightMarketInsight object. For stream: true, the API may emit an event: quan_light_market_insight SSE event before completion. The metadata is omitted when Quan cannot ground it to real source-linked quote context.

{
  "model": "quan-3.4",
  "messages": [
    { "role": "user", "content": "How is AAPL performing today?" }
  ],
  "stream": false,
  "includeInsightMetadata": true
}

JavaScript Example

This is the basic pattern for a Node.js backend, serverless function, or agent tool wrapper. Keep the API key on the server side rather than shipping it to a browser client.

const response = await fetch("https://stockup.cc/v1/query", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "x-api-key": process.env.STOCKUP_API_KEY
  },
  body: JSON.stringify({
    model: "quan-3.4",
    messages: [
      {
        role: "user",
        content: "Why is NVDA moving today? Give confirmed facts, possible drivers, risks, and what to watch next."
      }
    ],
    stream: false,
    webSearch: true,
    temperature: 0.2
  })
});

const data = await response.json();
const answer = data?.candidates?.[0]?.content?.parts?.[0]?.text;
console.log(answer);

Python Example

Python works well for research notebooks, batch portfolio checks, Slack bots, and internal analyst tools.

import os
import requests

payload = {
    "model": "quan-3.4-deep-research",
    "messages": [
        {
            "role": "user",
            "content": "Audit Tesla's latest risks, margin pressure, demand signals, and valuation setup. Cite uncertainty clearly."
        }
    ],
    "stream": False,
    "webSearch": True,
    "temperature": 0.2
}

res = requests.post(
    "https://stockup.cc/v1/query",
    headers={
        "Content-Type": "application/json",
        "x-api-key": os.environ["STOCKUP_API_KEY"]
    },
    json=payload,
    timeout=120
)

res.raise_for_status()
data = res.json()
print(data["candidates"][0]["content"]["parts"][0]["text"])

Streaming Responses

Set stream to true when you are building a chat UI, research workspace, or AI agent that should show partial output while Quan is reasoning. The API returns server-sent events.

curl -N -X POST https://stockup.cc/v1/query \
  -H "Content-Type: application/json" \
  -H "x-api-key: STOCKUP_API_KEY" \
  -d '{
    "model": "quan-3.4",
    "messages": [
      {
        "role": "user",
        "content": "Build a concise bullish and bearish thesis for AMZN using current market context."
      }
    ],
    "stream": true,
    "webSearch": true
  }'

Deep Research Prompt Pattern

For quan-3.4-deep-research, give the model a clear research job, ticker scope, output shape, and uncertainty standard. That helps AI agents produce repeatable financial research rather than loose commentary.

{
  "model": "quan-3.4-deep-research",
  "messages": [
    {
      "role": "user",
      "content": "Create a deep research memo on GOOGL. Cover business segments, valuation, margin trend, AI capex risk, regulatory risk, recent catalysts, bear case, bull case, and 5 questions an analyst should ask next. Separate facts from judgment."
    }
  ],
  "stream": false,
  "webSearch": true,
  "temperature": 0.15
}

OpenAPI and MCP for AI Agents

Quan 3.4 is designed to be discovered and called by AI agents. Use the public OpenAPI YAML or OpenAPI JSON for custom GPT Actions, internal workflow builders, and tool-calling agents. Use the StockUp MCP guide or the public MCP server script to connect Quan to Claude Desktop, Cursor, and other MCP-compatible clients.

• Custom GPT financial action: import https://stockup.cc/openapi.yaml and authenticate with your StockUp API key.
• Claude financial MCP: run the StockUp MCP server with STOCKUP_API_KEY in your client environment.
• Cursor finance agent: expose the same MCP tools for stock comparison, SEC filing review, sentiment analysis, DCA backtesting, and portfolio risk.

Related Financial AI Resources

Choose the guide that matches the product you are building:

• Financial AI API capabilities and pricing
• Stock market AI API workflows
• SEC filing AI reader workflows
• Quote-data-plus-reasoning alternative
• MCP financial tools for compatible AI agents

Production Checklist

1. Store STOCKUP_API_KEY in server-side environment variables or a secret manager.
2. Use quan-3.4 as the default model for most user-facing queries.
3. Route long-form SEC, portfolio, or due-diligence tasks to quan-3.4-deep-research.
4. Set stream: true for chat interfaces and stream: false for jobs that need one JSON payload.
5. Ask prompts to separate confirmed facts, model judgment, and unknowns for cleaner financial outputs.
6. Link agents to llms.txt, OpenAPI, and MCP so tools can discover StockUp reliably.