Niches · Financial market data
Build a local dataset for research or backtests
One company's docs show a full fit.
What counts as done
The full universe and range land in the store from bulk files or batched jobs, with the cost known before the download and symbols mapped across renames.
Measured, for our own tests: cost estimate within 10 per cent of the charge. No docs we read state these figures, so they don't count against a listing here.
How agents ask for it: download ten years of minute bars for the S&P 500
, load options history into my database
, backtest this strategy on all US stocks
Does the job 1
The docs show it meets the test above, and nothing we read says otherwise.
| Listing | Why |
|---|---|
| Databento C Databento, Inc. | metadata.get_cost returns the US dollar cost of a request and get_billable_size its bytes before download; batch downloads deliver large ranges as files; stype_in parent and continuous plus the security master map symbols. Duplicate streaming requests are charged twice. source |
Does the job, with a limit 7
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| Alpha Vantage E Alpha Vantage Inc. | No bulk files or batch jobs; a universe is fetched symbol by symbol, intraday one month per call, at 75 to 1,200 requests a minute. Cost is a flat plan price, and LISTING_STATUS with date and state=delisted helps survivorship. source |
| EODHD C Unicorn Data Services | GET /eod-bulk-last-day/{EXCHANGE} returns an exchange's day, with splits and dividends, for 100 calls; costs per endpoint are published and /user reports usage. Long ranges still mean one call per ticker, and rename history covers US only since July 2022. source |
| Eulerpool Financial Data API D Eulerpool Research Systems Pte. Ltd. | Dataset Snapshot pages per-ticker rows (limit up to 2,000) with as_of for market-wide dumps, and an EOD bulk exists; no cost estimate endpoint or symbol-rename mapping was found. source |
| Intrinio C Intrinio, Inc. | GET /bulk_downloads/links returns the account's bulk files with format and byte length; which datasets and their cost come with the plan or an order form, with no cost estimate per request. source |
| Massive (formerly Polygon.io) BB Massive.com, Inc. | S3 flat files deliver day and minute aggregates, trades and quotes for all US equities as daily unadjusted CSV, with Ticker Events for renames, and acc-06 loaded 450 GB; which plan includes flat files, and so the cost, was not read here. source |
| Tiingo D Tiingo Inc. | Flat plan prices make cost known, CSV cuts bandwidth and supported_tickers.zip maps symbols; but prices come one ticker per request against hourly, daily and monthly bandwidth caps, and delisted price coverage starts about 2014. source |
| Twelve Data BB Twelve Data Pte. Ltd. | /batch submits many requests asynchronously with credits summed per request, and returns partial data once credits run out; there are no bulk files and no cost estimate before a job. source |
Couldn't check 1
The docs were closed to us, or silent. These carry no mark either way.
| Listing | Why |
|---|---|
| Financial Modeling Prep D Financial Modeling Prep | site.financialmodelingprep.com answered 403 to robots.txt, so no FMP docs were opened; no account covers bulk downloads. |
Doesn't do this job 3
Listings in the category that do something else, with the reason.
| Listing | Why |
|---|---|
| Benzinga D Benzinga | No flat files or batch jobs for prices; bars are an on-request endpoint and no per-request cost or estimate is published. source |
| Financial Datasets C Financial Datasets, Inc. | No bulk files or batch jobs; per-call pricing is known, but a universe means one paged request per ticker, and the market snapshot is latest only. source |
| Finnhub D Finnhub | No bulk price files or batch jobs; intraday candles come one month per call per symbol, and only SEC filings and financials point to Kaggle dumps. No cost estimate. source |
2 listings in the category were added after this reading and aren't marked yet.
What stops agents finishing
- per-GB price visible only behind a login
- duplicate requests charged twice
- survivorship and look-ahead bias the data cannot flag
Around it
Before: chat.user-strategy. After: code-sandboxes.run-backtest, data.query-sql.
Where this comes from
Weight 10 of 100 in Financial market data: 5 of 46 accounts (5 of 48 mentions); Massive and Databento users building local stores, plus the point-in-time request
The same as data: /api/v1/use-cases/market-data.research-dataset.json, or the MCP tool get_use_case with id market-data.research-dataset. Other jobs in this category: /api/v1/niches?category=market-data.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.