Explain a market move
Combine server-verified quote context with possible drivers, source-aware research, uncertainty, and what to watch next.
Financial AI API · built for production workflows
Finance answers your app can trace.
Quan gives your product a single endpoint for stock research, SEC filing analysis, portfolio risk, market-move explanations, and cited web research—without making you stitch a generic model to a pile of finance prompts.
- 100,000 Quan 3.4 L tokens
- No card to start
- Two searches included per grounded request
first-request.sh
curl https://stockup.cc/v1/query \
-H "x-api-key: $STOCKUP_API_KEY" \
-H "content-type: application/json" \
-d '{
"model": "quan-3.4",
"messages": [{
"role": "user",
"content": "Compare AAPL and MSFT.\nSeparate verified facts from interpretation."
}],
"webSearch": true,
"stream": false
}'
Server-side use only · keys begin with sk_quan_
1 endpointOpenAI-style messages
4 public aliasesLite through Deep Research
8 MCP toolsFor Claude, Cursor & agents
$10 minimumPrepaid premium access
What you can ship
Useful financial workflows, not a generic chat wrapper.
Review a 10-K or 10-Q
Surface changes in risk factors, liabilities, guidance, working capital, margins, and earnings quality. See the workflow.
Find concentration risk
Review holdings for overlap, factor exposure, drawdown risk, catalysts, and concrete research priorities.
Compare stocks consistently
Ask for valuation, growth, profitability, sentiment, catalysts, risks, and evidence in one repeatable request.
Add finance tools through MCP
Install a dependency-free server that exposes quote, filing, sentiment, DCA, portfolio, comparison, and market-move tools.
Inspect the reasoning payload
Relevant Quan 3.4 tasks can return decision IDs, task intent, calculation and evidence metadata, release fingerprints, and quantitative-core hashes alongside the answer.
From zero to first call
Three steps. No sales call.
Verify your email
Create a developer account and verify the address before generating a key.
Run Quan 3.4 L free
Use the included 100,000 tokens to test the endpoint, streaming, and grounded prompts.
Continue when it works
Add $10, $25, or $50 for token overage, more research, or deeper models. Paid calls stop at zero.
Choose the right model
Spend the reasoning budget where it matters.
Prototype cheaply with Quan Lite. Move core analysis to Quan 3.4. Reserve Deep Research for longer, multi-step work.
- Quan Lite: fast briefs, classification, iteration, and agent routing.
- Quan 3.4: flagship analysis, comparisons, filings, catalysts, and grounded research.
- Deep Research: portfolio reviews, filing audits, and bounded multi-stage synthesis.
Purpose-built endpoints
Beyond /v1/query.
Most work starts at the messages endpoint, but seven quantitative endpoints sit alongside it. They accept structured input, return structured output, and — unlike a model call — the figures they produce are computed rather than generated. Each has its own guide.
DCF, WACC & Monte Carlo
Deterministic valuation with a seeded simulation. Missing inputs come back as unavailableInputs rather than as an invented default.
Portfolio risk
VaR, CVaR, Sharpe, Sortino, max drawdown and beta from holdings alone, plus checks against five constraint limits.
Macro regime
Seven-state classification with probabilities, transition risk and an explicit state for when the signals genuinely conflict.
Walk-forward backtest
Choose a date window and a transaction cost. Costs default to 10 basis points rather than to zero.
Composite signal
A provenance-aware reading that reports its own coverage, so thin support stays visible in your interface.
Decision intelligence
Regime, risk, calculations and evidence composed into one auditable response with a claim ledger.
PDF extraction
Resumable upload to 50 MB, OCR with per-page unreadable reporting, then query the document by ID.
Decision audit
Every quantitative response carries a decisionId. Pass it here for the owner-scoped immutable record, months later.
Questions developers ask
Know what you are buying.
Is this a stock-data API or an AI reasoning API?
Quan is primarily a financial reasoning API. It can enforce server-verified current quote context for supported stock requests, but it is not a drop-in replacement for every raw market-data endpoint, exchange feed, or historical OHLCV database.
What does “verified” mean here?
For supported quote requests, StockUp retrieves quote context server-side and instructs the model to treat that block as authoritative. If a quote is unavailable, the model is required to say so instead of estimating an exact current price. You should still review primary sources for consequential decisions.
Can I use Quan with Claude Desktop or Cursor?
Yes. The StockUp MCP server supports compatible clients and exposes eight finance-specific tools. You can install it with npx -y @stockup/quan-mcp-server.
Will I get a surprise bill?
No. Self-service access is prepaid. StockUp reserves estimated request cost, settles actual usage, and pauses paid calls when your wallet cannot cover the request.
The first calls are on us
See if Quan earns a place in your stack.
Create a key, run Quan 3.4 L against a real workflow, and add money only when you need more than the included allowance.
Start with 100,000 tokens →