MarketParquet vs Finnhub
> data API strong on fundamentals + alternative data
Finnhub is a broad data API with a real strength in fundamentals and alternative data -- news sentiment, filings, insider and congressional trades, ESG, transcripts. But it moved US stock historical candles behind a paid plan around 2025, so the free tier no longer returns OHLCV. MarketParquet does one thing: bulk US OHLCV history as Parquet, one-time. Here's the honest comparison.
Side-by-side
| Finnhub | MarketParquet | |
|---|---|---|
| Asset classes | US + global stocks, ETFs, forex, crypto, + deep fundamentals & alt-data | US stocks, ETFs, futures |
| Historical OHLCV on free tier | no -- US stock candles are paid (removed from free ~2025) | free recent daily sample; full archive one-time |
| Format | JSON via REST + WebSocket | Apache Parquet, by-date partitioned |
| Finest resolution | 1-minute (tick on higher tiers) | 1-minute bars |
| Real-time data | yes (WebSocket) | no -- end-of-day batch only |
| Pricing | modular ~$50/mo per data category; stacks to ~$150-200+ | one-time $79 Complete Archive |
| Split / dividend adjusted | split-adjusted; dividends are a paid add-on | yes (split-adjusted) |
| Alt-data / fundamentals | yes -- news sentiment, filings, insider/congressional, ESG | no -- OHLCV only |
| Includes delisted tickers | not a documented feature | yes -- included |
Where Finnhub wins
- Unusually broad alternative data -- news sentiment, SEC filings, insider and US congressional trading, ESG, and earnings-call transcripts.
- Deep global fundamentals and analyst estimates in the same API.
- Modern, well-documented REST + WebSocket API with real-time streaming and (on higher tiers) tick history going back decades.
- Low entry cost and a usable free tier for non-commercial prototyping.
Where MarketParquet wins
- Full US OHLCV history as bulk Parquet, one-time -- Finnhub moved US stock historical candles behind a paid plan (removed from free ~2025).
- No per-call rate limits: load a whole trading day at once instead of paging a REST API.
- One-time $79 vs Finnhub's modular ~$50/month-per-category pricing that stacks up quickly.
- Delisted tickers included for survivorship-bias-free backtests; ready split-adjusted bars (Finnhub gates dividends behind paid endpoints).
Which to pick
Pick Finnhub if you want real-time quotes, global fundamentals, or alternative data like news sentiment, filings, and insider/congressional trades.
Pick MarketParquet if you need bulk US OHLCV history for backtesting, want Parquet files you keep, and don't want historical candles behind a subscription or per-call limits.
FAQ
Is Finnhub still free for historical stock data?
Not for US stock candles. Finnhub moved US historical OHLCV candles (and dividends) to paid plans around 2025 -- the free tier returns an access error on the stock-candle endpoint. MarketParquet gives you a free recent-daily sample and the complete US archive as Parquet for a one-time $79.
Does Finnhub provide bulk or Parquet downloads?
No -- Finnhub is an API-first service (JSON over REST plus WebSocket), so you request data symbol by symbol and there's no bulk flat-file or Parquet export. MarketParquet ships Parquet partitioned by trading day, so one file loads a whole day of every ticker.
Finnhub vs MarketParquet -- which for backtesting?
For bulk US equity/ETF/futures backtesting on local files, MarketParquet is simpler and cheaper: Parquet you keep, no per-call limits, one-time $79. Finnhub is better if you also want real-time data, global fundamentals, or its alternative-data feeds (news sentiment, filings, insider and congressional trades).
Is Finnhub's data split- and dividend-adjusted?
Finnhub's daily candles can be split-adjusted via a parameter, but dividend adjustment isn't applied by default and the dividends endpoint is a paid feature. MarketParquet ships split-adjusted OHLCV out of the box, with delisted tickers included.
> Pricing and feature snapshots reflect public information at time of writing and may shift. Always confirm against Finnhub's own pricing page before committing.