REST + MCP - Now Available

Seasonal patterns and ML scores for US markets, in one API call

Query 1-99 years of seasonal history - by calendar or election cycle - with an ML win-probability on every setup, over REST or straight from ChatGPT and Claude via MCP. Try it in ~30 seconds below with a live demo call, then build, embed, or ship on it.

# Real authenticated call - no signup, no key, paste and run curl https://api.tradewave.ai/v1/scan?market=2&limit=5 \ -H "Authorization: Bearer tw_demo_explore"

tw_demo_explore is a public demo token - it is a constant, not a secret, so paste it as-is. Returns today's ranked seasonal patterns for AAPL, MSFT, NVDA, AMZN and TSLA. A free key unlocks every symbol.

No credit card. The Free tier is open now - paid plans (Dev, Pro, Business, and Enterprise with volume licensing, an SLA, and a signed agreement) launch soon - talk to sales.

Derived research, not a data firehose

Raw market data is a commodity; execution needs capital and a license. The scarce, defensible layer is the one in between - the quantified reason a setup is worth a look. Each detected seasonal pattern ships as percentages, a 0-100 seasonal index and an ML win-probability, with the dates that bound the window - never raw OHLCV. Because we return derived values, not exchange-licensed quotes, there are no exchange agreements, per-terminal fees, or market-data audits on your side.

🎭

ML win-probability on seasonal patterns

A 62-feature model trained on millions of historical setups scores every seasonal opportunity with a win probability and predicted return before market open. Every tier includes ML scores (free starts at 5/day; Pro is unlimited). No other seasonality API offers it.

📈

Verifiable, time-stamped track record

Every daily pick is recorded before it opens. The get_pick_track_record endpoint returns the realized win/loss history - not a backtested curve, a forward-looking ledger you can audit call by call.

🤖

Agent-native via MCP

The TradeWave MCP server is a hosted HTTP endpoint: sign in from ChatGPT or Claude.ai with your TradeWave account, or connect Claude Desktop, Cursor, or any MCP-compatible host with an API key. Ask your AI assistant to find the strongest seasonal longs in S&P 500 stocks, rank by ML score, and compare to the live track record - no glue code required. TradeWave supplies the seasonal and ML edge plus the timing; your assistant pairs it with its own fundamentals, news, and macro tools.

Where TradeWave sits

We are the layer above the trade, not another place to make it. Use TradeWave for the reason and the timing, then render it in your own app, route it to your agent, or hand the ticket to whatever you already execute with - any in-chat execution app or broker.

Raw data feeds

Prices, quotes, and history sold by the call. A commodity in a race to zero - they hand you numbers but no edge and no timing, so the decision is still entirely on you.

Execution apps and brokers

The last mile that places the order and clears the trade. Most never publish a forward-tested track record, and a venue that earns per trade is structurally conflicted, so "sit this one out" is not something they are built to tell you.

TradeWave - the research layer

The quantified reason and the timing, as a Pattern Card with verifiable, time-stamped, forward-tested receipts - and an honest empty result when nothing scores. We sit between the other two: derived values out (percentages and a 0-100 seasonal index), no raw prices in your response.

Your agent can write a backtest. It cannot make it true.

A capable agent can scaffold a seasonality backtest in minutes. What it cannot conjure is the licensed data, the trained model, and the forward-tested receipts that decide whether the backtest is true or just confidently wrong. Consume that edge in one call instead of rebuilding it badly.

For AI agents - why consume, not compute

Audited, not believed

We do not show you a curve to admire. Every daily pick is logged before the outcome, then scored against what actually happened - a forward record you can audit call by call, not a backtest fit after the fact. Read the live record instead of taking our word for it.

Year Base seasonal win rate ML-selected picks win rate ML lift Picks
Example 64% 79% +15pp 57

One example row, shown only to explain how the ML layer is meant to lift the base seasonal win rate - it is a concept, not a result. The real numbers are public and call-by-call: pull the realized record from /v1/daily-pick/track-record on the free tier, or read the same record on the public scorecard. Past performance does not guarantee future results.

See the Public Record Pull It via the API

What Makes TradeWave Different

TradeWave is the only seasonality engine that both shows its work and scores the odds: an AI-calibrated win-probability on every pattern, a forward-recorded track record, and native REST and MCP access - at a fraction of the cost of a data terminal, open to any developer or AI agent today.

Feature TradeWave API Typical seasonality tools Data terminals
ML win-probability scoringAll tiers (unlimited on Pro+)NoNo
REST API accessYes - all tiersRareTerminal only
MCP / AI agent integrationYes - nativeNoNo
Verified pick track recordYes - forward-recordedBacktest onlyNo
Redistributable to your own usersYes - by designMixedRaw data only
No data terminal or seat requiredYesVariesRequired
Starting priceFreeHundreds per monthTens of thousands per year

Try, build, ship

Start free, scale when it earns its place. The full ladder is on the pricing page.