Clear guides to crypto, markets and the technology behind them.
Balancer holders approved winding the protocol down, with 99.2% of 17,226,164 BAL in favour. Pools move to withdrawals only on 30 October 2026, and the treasury returns to holders from May 2027.
Curve publishes 182 API endpoints across two keyless services. On 23 September 2026 neither recognized Plasma, where DexPaprika measured $1.78m of daily Curve volume.
Uniswap (UNI) closed at six different prices in one hour on 29 September 2026, one per pool. Here is why pool candles disagree, and how token OHLCV turns them into one VWAP series.
On 23 September 2026 LINK showed 70 different prices across DEX pools on eight chains. The liquid pools agreed within 1.4%, and their median matched CoinPaprika's price to a tenth of a cent.
On 17 September 2026 a pool seeded with about 1.5 SolvBTC pushed NSTR's oracle price from $0.006 to $49.5, and one account borrowed about $3.5 million from Nostra. Here is how the feed was fooled, and what to check in a pool first.
Orca runs a free, keyless public API for its Solana Whirlpools. On 23 September 2026 the protocol carried 8.1% of DexPaprika-indexed Solana DEX volume, and its own API has no price candles at all.
Meteora runs five separate public APIs, and two of them return real OHLCV candles without a key. On 23 September 2026 its pools carried 15.7% of Solana's DexPaprika-indexed DEX volume, yet none of the five has a swap feed.
At 22:17 UTC on 23 September 2026 the biggest row on Ethereum was a MakerDAO stablecoin converter, at 37.9% of indexed volume. Here is every chain's leading DEX, its share, the chain total, and the anomalies to check first.
DexScreener documents 300 calls a minute on its pair endpoints and 60 on the rest, then sends nothing back to pace against. Here is what 36 requests on 23 September 2026 showed about its cache.
Pump.fun publishes two Solana programs and an SDK, not a REST API. On 23 September 2026, Pump.fun and PumpSwap together carried 69.65% of Solana DEX volume in DexPaprika's index.
Balancer's GraphQL API covers its own pools in full. According to DexPaprika data, Balancer held about 4.3% of Polygon's DEX volume on 23 September 2026, two days before a vote on shutting the protocol down.
PancakeSwap carried 91.8% of BNB Smart Chain DEX volume on the day I measured it, and documents no REST API at all. Here is what its docs really offer, and why the top Google result leads to a 500.
Raydium runs a real public API and it needs no key at all. It also has no price candles and no swap feed, and it covers a protocol that was 5% of Solana DEX volume on the day I measured it.
On 22 September 2026 Uniswap v4 moved more Ethereum volume than v3, yet the Uniswap API returns no pool reserves, no history and no candles. Here is what it covers, where Uniswap points you for data, and what that costs.
Use this to match a crypto API to the job: CoinPaprika for free, commercial-grade market data with no key, CoinStats for wallet and portfolio apps, and DexPaprika for real-time DEX and on-chain streaming — five providers compared.
One call returns the buy/sell split of a pool's flow across six timeframes. Learn the two ratios that measure pressure, why the divergence between them is the real signal, and the counts-vs-wallets caveat.
One GET request turns any contract address into a live USD price with liquidity, volume, and buy/sell flow attached, on 35+ chains. Plus the CoinPaprika cross-check route and a batched endpoint for watchlists.
Point Claude, ChatGPT, or Cursor at one hosted MCP URL and your assistant answers with live prices for 12,000+ cryptocurrencies instead of stale training data. Configs for every client, plus the JSON-RPC route for coded agents.
One HTTPS call gets you a year of daily prices, volume, and market cap for 2,000 coins, no API key. Here's exactly how deep free historical crypto data goes, where candles start costing money, and the code to pull it all.
A liquidity drain is how most rug pulls take the money, and it shows up as one strongly negative reserve change. Learn the two-threshold rule that catches it in the same block, plus the JIT traps that cause false alerts.
