API reference

DataCedar serves historical price bars for listed equities, ETFs and headline indices on every exchange it covers. Daily history runs back to each listing's first day. One minute bars are captured every session from the day capture began for an exchange and are never discarded; the coarser intraday frames (5m, 15m, 1h) are rolled up from those minutes on read. Daily bars are fetched separately from the upstream and may differ in prices and volume. Intraday coverage can omit auction executions.

Everything is organised by exchange and country. One registry lists the exchanges; every symbol id names its exchange; every list endpoint filters by exchange or country. The same data is on every symbol page, on the REST API documented here, and in the MCP server for agents.

Create a key in the dashboard, then pull a year of daily bars for one symbol.

curl -s "https://api.datacedar.com/v3/stocks/nasdaq:AAPL/bars?timeframe=1d&from=2025-09-01&order=asc" \
  -H "Authorization: Bearer $DATACEDAR_KEY"

{
  "data": [
    { "sid": "nasdaq:AAPL", "symbol": "AAPL", "timeframe": "1Day", "ts_ms": "1756728000000",
      "open": 232.51, "high": 235.2, "low": 231.9, "close": 234.07, "volume": "41230311", "vwap": 233.8,
      "trade_count": "0", "adjustment": "raw" },
    ...
  ],
  "pagination": { "count": 261, "has_more": false, "limit": 500, "next_cursor": null },
  "meta": { "request_id": "req_…", "timestamp": "…", "sid": "nasdaq:AAPL", "exchange": "nasdaq",
            "currency": "USD", "timezone": "America/New_York", "attribution": "…" }
}

Do not know the id? GET /v3/search?q=apple returns it. Want to see what is covered? GET /v3/exchanges and GET /v3/countries.

Every request carries your key, either as Authorization: Bearer <key> or as X-API-Key: <key>. Keys are created and revoked in the dashboard; a key belongs to an account and inherits the account's plan. Keep keys server side; the browser never needs one because every symbol page is public.

Base URL https://api.datacedar.com/v3. Field additions are not versioned; a breaking change to a field's meaning or type ships under a new path version with the old one kept alive for twelve months.GET /v3/health needs no key.

A symbol id (sid) is {exchange}:{TICKER}: nasdaq:AAPL, nyse:IBM, nyse-arca:SPY. The exchange code is the one on the markets page and in /v3/exchanges; the ticker is the exchange's own. Matching is case insensitive on input; responses always return the canonical form. The same company on two exchanges is two ids, on purpose: they trade in different currencies on different calendars.

A bare US ticker (AAPL) is accepted too and resolved to its exchange, so nothing written against the first version of the API breaks. Every asset payload carries its sid, the bars_url to call next and the page it lives on.

Every endpoint returns the same envelope: data, pagination, meta.

Errors are { "error": { "code", "message", "details", "request_id" } }. The message says what to do next.

Every response carries X-RateLimit-Limit, X-RateLimit-Remaining and, when throttled, Retry-After.

Bar endpoints take timeframe (1m 5m 15m 1h 1d), from and to (YYYY-MM-DD, ISO 8601 or epoch milliseconds), adjustment (daily bars only), order (desc default, asc), limit (default 500, max 50,000 daily and 5,000 intraday) and cursor. When a response is cut, pagination.next_cursor resumes exactly after the last bar returned, with no overlap and no gap.

List endpoints (/v3/exchanges, /v3/stocks/assets, /v3/search) take exchange and country. Both accept a comma list; country takes an ISO code (US), a slug (united-states) or a name.

With no from, a request starts at the edge of your window. A from beyond it answers 403 HISTORY_LIMIT_EXCEEDED rather than a truncated series, so a study never silently loses its tail. See pricing.

Every exchange in the registry: code, name, country, currency, timezone, indices, languages, whether it is live and how many symbols it has. Filters: country, region, live, wave, language.

curl -s "https://api.datacedar.com/v3/exchanges?live=true" -H "X-API-Key: $DATACEDAR_KEY"

One exchange, plus a count of its symbols by asset class.

The exchange's symbols, alphabetical, paged (limit up to 1,000, cursor). Filters: letter, asset_class (us_equity, equity, etf, index), q.

Countries with their exchanges, grouped from the same registry. Filters: region, live.

Find a symbol by ticker or name. Exact ticker first, then ticker prefix, then name. Filters: exchange, country, limit (max 50).

curl -s "https://api.datacedar.com/v3/search?q=apple&country=US" -H "X-API-Key: $DATACEDAR_KEY"

The master record for one symbol: name, exchange, asset class, status, bars_url, page.

The whole master, paged 1,000 at a time. Filters: exchange, country, asset_class, status, q.

Cash dividends and splits, oldest first. Filters: type (dividend, split, all), from, to as YYYY-MM-DD.

Bars for one symbol. timeframe, from, to, adjustment, order, limit, cursor.

What is stored for one symbol: per timeframe the first bar, the last bar and the bar count, in one call. Use it before a history pull to size the request, or to state coverage.

Bars for several symbols: symbols=comma separated ids, up to the plan's batch size, same parameters. data is a map keyed by sid. Every series is cut at the same instant and one next_cursor continues all of them together, so a page is directly comparable across symbols.

import requests

H = {"X-API-Key": KEY}
r = requests.get(
    "https://api.datacedar.com/v3/stocks/bars",
    params={"symbols": "nasdaq:AAPL,nasdaq:MSFT,nyse:IBM", "timeframe": "1d", "from": "2025-01-01", "adjustment": "all", "order": "asc"},
    headers=H,
).json()

for sid, bars in r["data"].items():
    print(sid, bars[0]["ts_ms"], "->", bars[-1]["ts_ms"], len(bars), "bars")

Exchanges come online in launch order, published on the markets page and on /v3/exchanges, where liveflips the day an exchange's symbols have bars. When an exchange comes online, every listed symbol gets its full daily history and minute capture starts the same day. Minute bars are captured every session and never discarded.

Daily bars are stored raw and adjusted; adjustment=all applies the cumulative split and dividend factor to every bar before the action date, split and dividend apply one family each, and raw applies nothing. Intraday frames are rolled up from stored one minute bars on read; the roll up is deterministic, so two requests for the same range return identical series.

The MCP server exposes the same data as the API, through the same handlers, so filters and plan windows behave identically: list_exchanges, list_countries, find_symbol, get_symbol, list_symbols, get_bars, get_actions, whoami. Authenticate with your key.

claude mcp add --transport http datacedar https://api.datacedar.com/mcp --header "Authorization: Bearer $DATACEDAR_KEY"

A machine readable summary of the site and the API is published at /llms.txt.

Newest first:

• 2026-09-28. Exchange registry: /v3/exchanges, /v3/countries, /v3/search, exchange qualified ids, exchange and country filters, batch bars, corporate actions, plan windows enforced, MCP stocks tools.
• v2. DataCedar rebuilt around one minute bars for every exchange: symbol pages, exchange hubs, MCP.

Minute history fills forward from capture start and cannot be backfilled; it matures toward your plan window over calendar time. Questions? Get in touch.