patternfetch Stock technical indicators API
Stock indicators · API + MCPRSI, EMA and ATR on stocks — as a label, on adjusted prices.
One POST /v1/brief with a stock or ETF ticker returns the interpreted state of the indicators, not a 200-number array: overbought, above_20_50, a volatility percentage — with the exact numeric value still attached. And because equities split and pay dividends, everything is computed on adjusted candles, so a corporate action never fakes a crossover your agent then acts on.
StocksETFsRSI · EMA · ATR1m → 1wYahoo · adjusted
TL;DR — One POST /v1/brief (or the MCP tool patternfetch_brief) with a stock or ETF ticker returns interpreted indicator state: RSI as { v, state: overbought | oversold | neutral }, EMA as { v, state: above_20_50 | below_20_50 | mixed | above_20 | below_20 }, and volatility as volPct (ATR as a percentage of price, in the regime block). All computed on split- and dividend-adjusted candles. The same call also returns patterns, support/resistance levels and a one-line nl summary.
Why a label instead of an array
An LLM reasoning over a long array of indicator numbers does two things badly: it burns tokens carrying the array, and it hallucinates when it has to compare, average or cross those numbers itself. A state label sidesteps both. overbought is cheap to carry and unambiguous to act on; above_20_50 already encodes the moving-average relationship the agent would otherwise derive by hand — badly. The numeric v is still in the payload for anyone who needs the exact figure, so you lose nothing; you just stop paying for arithmetic the model gets wrong. Thresholds and windows are documented in the methodology. The failure mode itself is worked through on the indicator hallucination page.
Why adjusted prices decide the answer
This is the part that has no crypto equivalent. A crypto pair trades continuously and never splits; a stock does both, and an indicator computed on the unadjusted series will confidently report something that never economically happened.
What breaks unadjusted
- A 4:1 split drops the raw price 75% overnight. RSI reads that as a violent selloff and can print oversold on a day nothing happened.
- A dividend knocks the price down by the payout — small, but it accumulates across the EMA window.
- Either can push price across its 20 or 50 EMA and manufacture a crossover out of an accounting event.
What patternfetch does
- Indicators are computed on split- and dividend-adjusted candles from Yahoo.
- The state label therefore describes the market, not the corporate action.
- The same adjusted series feeds patterns and levels, so every block of the brief agrees with every other block.
The one call
A real /v1/brief shape for Apple on the daily. Indicators arrive interpreted; volPct rides along in regime.
Request
curl -X POST https://patternfetch.com/v1/brief \
-H "Authorization: Bearer pf_…" \
-H "Content-Type: application/json" \
-d '{ "ticker": "AAPL",
"timeframe": "1d" }'Response (indicators, interpreted)
"indicators": {
"rsi": { "v": 61.4, "state": "neutral" },
"ema": { "v": 229.83, "state": "above_20_50" }
},
"regime": {
"trend": "up",
"strength": 0.38,
"volPct": 1.42
}The same call also returns detected chart & candlestick patterns, clustered support/resistance levels and a one-line nl summary — the indicators above are one block of the full brief, not a separate endpoint to stitch in. See it live →
The state vocabulary
A small, closed set of labels — which is the point. An agent can branch on these without a parser and without a prompt explaining what 71.2 means.
| Field | States | What it encodes |
|---|---|---|
| rsi.state | overbought · neutral · oversold | Momentum extreme, from the RSI value |
| ema.state | above_20_50 · above_20 · mixed · below_20 · below_20_50 | Price position against the 20 and 50 averages |
| regime.trend | up · down · sideways | Directional regime over the window |
| regime.volPct | numeric % | ATR as a percentage of price — includes overnight gaps |
On equities volPct carries something a continuous market's doesn't: because ATR uses true range, the gap between one session's close and the next session's open is inside the number. For an agent sizing overnight exposure that is the correct reading — the gap is the risk. Methodology →
Free to start. MCP-native for agents.
Streamable HTTP at /mcp — discovery is free, a tools/call is billed like REST. Five tools, one schema across stocks, ETFs and crypto.
Connect (OAuth or Bearer)
# one line — OAuth mints a free key,
# nothing to paste
claude mcp add --transport http \
patternfetch https://patternfetch.com/mcp
# or Bearer in a config file
{ "mcpServers": { "patternfetch": {
"url": "https://patternfetch.com/mcp" } } }Free, then pay-per-call
Free to start: ① No-signup demo — POST /v1/demo returns a real brief with indicators, no key. ② Free key with $3.00 starter credit (300 briefs) from one call, no card. A brief is $0.01. Pay with x402 (USDC on Base, no account) or Stripe. Credits never expire. Pricing →