Technical indicators: RSI, EMA, MACD, Bollinger Bands and VWAP

What you’ll build

A screen that reads RSI and MACD for a pool straight from the API, without downloading candles and computing the indicator yourself. The server computes it from the same OHLCV candles that pool OHLCV returns, so the numbers line up with the charts you already draw. By the end you will have:

A curl that returns hourly RSI for a pool, with no key
A Python script that flags overbought and oversold pools and MACD crosses
A clear picture of which indicator needs which plan

Which plan you need

	Keyless and free key	Dev, Pro, Enterprise
Pool RSI, EMA, MACD	Yes, on candles of `10m` and up and the last 7 days	Every interval, no history limit
Pool Bollinger Bands, VWAP	No, `403`	Yes
Token RSI, EMA, MACD, Bollinger Bands, VWAP	No, `403`	Yes

Paid plans call https://api-pro.dexpaprika.com with the key as the entire Authorization header. Compare plans on pricing and get a key in console.dexpaprika.com.

The call

Hourly RSI for the USDC/WETH pool on Uniswap v3, last six hours, no key:

curl "https://api.dexpaprika.com/networks/ethereum/pools/0x88e6a0c2ddd26feeb64f039a2c41296fcb3f5640/indicators/rsi?interval=1h&start=-6h"

{
  "network": "ethereum",
  "pool_address": "0x88e6a0c2ddd26feeb64f039a2c41296fcb3f5640",
  "indicator": "rsi",
  "interval": "1h",
  "parameters": {"period": 14},
  "data": [
    {"timestamp": "2026-10-12T05:00:00Z", "value": 64.33426700982189},
    {"timestamp": "2026-10-12T06:00:00Z", "value": 66.54753114766311},
    {"timestamp": "2026-10-12T07:00:00Z", "value": 61.156181549269995},
    {"timestamp": "2026-10-12T08:00:00Z", "value": 54.96220386541894},
    {"timestamp": "2026-10-12T09:00:00Z", "value": 57.865403918793895}
  ],
  "has_next_page": false,
  "next_cursor": ""
}

Values come oldest first, one per candle. timestamp is the candle’s open time, and only completed candles count, so the candle that is still open has no value yet. parameters echoes what was used, defaults included. Set interval every time. It defaults to 24h, and on the free tier a daily series cut to the last 7 days is six values.

The five indicators

Replace rsi in the path with the indicator you want. Pools use /networks/{network}/pools/{pool_address}/indicators/{name}, tokens use /networks/{network}/tokens/{token_address}/indicators/{name}.

Indicator	Path	Parameters (default)	Each value carries
Relative Strength Index	`rsi`	`period` (14)	`value`, 0 to 100
Exponential moving average	`ema`	`period` (20)	`value`, on the same price as the candles
MACD	`macd`	`fast_period` (12), `slow_period` (26), `signal_period` (9)	`macd`, `signal`, `histogram`
Bollinger Bands	`bbands`	`period` (20), `std_dev` (2)	`upper`, `middle`, `lower`
Rolling VWAP	`vwap`	`period` (14)	`value`

Every indicator also takes:

Parameter	What it does
`interval`	`1m`, `5m`, `10m`, `15m`, `30m`, `1h`, `6h`, `12h`, `24h` (default) or `1d`, which is the same as `24h`
`limit`	Values per response, 1 to 1000, default 100
`start`	First candle: `-24h`, `-7d`, Unix seconds, RFC3339 or a date. Without it you get the last `limit` candles
`end`	End of the range, exclusive, same formats. Defaults to now
`cursor`	`next_cursor` from the previous page. Not combined with `start`

With start, limit counts forward from it. When the range holds more candles than limit, has_next_page is true and next_cursor continues it.

Screen a pool in Python

RSI over 70 is commonly read as overbought and under 30 as oversold. A MACD histogram that changes sign means MACD crossed its signal line. This script checks both on the last completed hourly candle:

import requests

API = "https://api.dexpaprika.com"


def indicator(network, pool, name, **params):
    r = requests.get(
        f"{API}/networks/{network}/pools/{pool}/indicators/{name}",
        params={"interval": "1h", "start": "-24h", **params},
        timeout=30,
    )
    r.raise_for_status()
    return r.json()["data"]


pool = "0x88e6a0c2ddd26feeb64f039a2c41296fcb3f5640"  # USDC/WETH, Uniswap v3

rsi = indicator("ethereum", pool, "rsi")[-1]
macd = indicator("ethereum", pool, "macd")[-2:]

state = "overbought" if rsi["value"] > 70 else "oversold" if rsi["value"] < 30 else "neutral"
print(f"RSI(14) at {rsi['timestamp']}: {rsi['value']:.1f} ({state})")

prev, last = macd
if prev["histogram"] < 0 <= last["histogram"]:
    print(f"MACD crossed above its signal line at {last['timestamp']}")
elif prev["histogram"] > 0 >= last["histogram"]:
    print(f"MACD crossed below its signal line at {last['timestamp']}")
else:
    print(f"No MACD cross in the last candle (histogram {last['histogram']:.2f})")

RSI(14) at 2026-10-12T09:00:00Z: 60.4 (neutral)
No MACD cross in the last candle (histogram 0.71)

Two requests, two credits. A screen over many pools reaches the keyless rate limit quickly; a free key raises it.

Four things that will surprise you

The same candle can get a different value from a different start. RSI, EMA and MACD are smoothed: each value carries some weight from every candle before it. The API reads only the candles each indicator strictly needs before your start, 14 for RSI(14), so a short range starts the smoothing late. Hourly RSI(14) for one 09:00 candle on the pool above, measured on 12 October:

`start`	RSI at 09:00
`-6h`	52.47
`-24h`	60.52
`-3d`	61.29
`-7d` and earlier	61.18, the same as from 30 days of candles

Start at least ten periods before the first value you use, 140 candles for RSI(14), and keep the tail. On the free tier the 7-day window holds 168 hourly candles, which is enough at 1h and shorter intervals, but only seven daily ones, so daily RSI and EMA need a paid plan to settle. The same effect explains differences from a charting site that has years of history behind each point. A short range on a quiet pool can fail. The warm-up is measured in time, period intervals back, so minutes without a trade leave it short of candles. interval=1m&limit=2 on the pool above answered 400 with insufficient data: rsi needs at least 15 candles. Received 12, and limit=50 worked. Ask for more values than you need. Quiet intervals have no candle, so they have no value. Indicators are computed over the candles that exist. On a thinly traded pool at 5m, the period of an RSI spans more wall-clock time than you would expect. Pool indicators follow the pool’s OHLCV price, with no way to flip it. An EMA, Bollinger Band or VWAP on a pool is on the same price as pool OHLCV with its default inversed=false, and the indicator endpoints take no inversed. For one USD series across every pool a token trades in, use the token indicators, built on token OHLCV.

Errors

Status	`error`	What happened
`403`	`plan_required`, no `param`	The endpoint needs Dev or higher: pool Bollinger Bands and VWAP, every token indicator
`403`	`plan_required`, `"param": "interval"`	An interval below `10m` without a paid plan
`403`	`plan_required`, `"param": "start"`	History before the last 7 days without a paid plan
`400`	`invalid_param`	A parameter out of range, named in `param`. For example `fast_period` must be below `slow_period`
`400`	`invalid_param`, no `param`	`insufficient data`: the range holds too few candles to compute the indicator. Widen it
`429`	`rate_limited`	Too many requests; wait `Retry-After` seconds

A paid-only request without a paid key looks like this:

{"error": "plan_required", "tier": "keyless", "message": "this endpoint requires a Dev or Pro plan", "required_tier": "dev"}

Where to go next

Indicators reference: every endpoint, parameter and response code
Pool OHLCV: the candles behind the pool indicators
Token OHLCV: the USD candles behind the token indicators
Pricing and console.dexpaprika.com: plans and your key

Get Started

Knowledge Base

Tutorials

SDKs

Tools

Company

Technical indicators: RSI, EMA, MACD, Bollinger Bands and VWAP

What you’ll build

Which plan you need

The call

The five indicators

Screen a pool in Python

Four things that will surprise you

Errors

Where to go next

​What you’ll build

​Which plan you need

​The call

​The five indicators

​Screen a pool in Python

​Four things that will surprise you

​Errors

​Where to go next

What you’ll build

Which plan you need

The call

The five indicators

Screen a pool in Python

Four things that will surprise you

Errors

Where to go next