The Angel One API — SmartAPI — occupies a specific niche among Indian data APIs: it’s free, it covers all three exchanges including MCX, and its login can run without a human. That last property makes the angel one api the favourite of unattended data pipelines — the collectors that must be streaming before the 9:15 open whether or not anyone’s awake. This guide examines it from the analytics consumer’s seat: the feed, the candles, the automation story, and the analytical layer that remains yours to build.
The feed and the fields
SmartAPI’s WebSocket 2.0 streams binary-packed ticks in modes from bare LTP to full quote with depth, across NSE, BSE and MCX — commodity derivatives included, which matters to anyone whose book extends past index options. Derivative packets carry open interest, the field every OI analytic keys on. Historical bars come from the getCandleData endpoint at one-minute resolutions upward with multi-year daily history. Instruments resolve from a large daily master JSON spanning every segment; index it locally, because it’s hundreds of thousands of rows.
The automation story: no browser in the loop
Where Zerodha and Upstox require a human to click a hosted login each morning, the angel one api authenticates with client code + PIN + TOTP in a single scripted POST — the TOTP seed enrolled once, generated programmatically forever after (the SmartAPI portal covers enrolment). Sessions still expire daily under exchange rules, but re-auth is one automated call. For a data collector that must never miss an open, this is the decisive feature, and it’s the reason so many home-grown OI trackers quietly run on SmartAPI.
What the raw feed doesn’t do for you
The recurring truth of every broker API applies here unchanged: ticks with OI are ingredients. The standard options-analytics stack — a coherent option chain assembled from hundreds of streams, OI-change baselines, buildup classification, IV solving, Greeks, participant and positioning reads — is all engineering the angel one api leaves to you. SmartAPI’s sparser documentation makes some of it slower going than on Zerodha or Upstox: the binary WebSocket layout in particular rewards patience and a packet dump.
Where SmartAPI fits best
Three profiles fit naturally. The unattended collector (its headline strength). The MCX-inclusive analyst, since free rivals like Groww don’t serve commodities at all. And the Angel One account holder who wants data without a second broker. If none of those is you, the API’s case weakens against Upstox’s cleaner keys and deeper candles — free either way, so the tie-breaks are ergonomic.
The shortcut for the standard layer
If the reason you’re evaluating the angel one api is to build the analytics everyone builds — live chain with OI analysis, Trending OI, screeners, PCR, max pain, GEX — that stack already runs at OIData on your own broker connection — you link your SEBI-registered account, market data streams straight from it, and there is no code to write or pipeline to babysit. The efficient split: let OIData carry the standard market-reading layer, and spend your SmartAPI effort only on computations genuinely yours — which is where any edge actually lives.
A note on data hygiene
One practical lesson from pipelines we’ve seen on SmartAPI: baseline discipline matters more than feed choice. OI analytics live and die on correct previous-session baselines per strike — get those wrong (missed contract rolls, stale masters) and every “OI up 18%” your system prints is fiction, whichever broker’s feed supplied the ticks. Whatever you build, test the baselines first.
A practical starter pipeline
The unattended-collector shape SmartAPI rewards: a scheduled job that generates the TOTP, posts the login and stores the day’s three tokens well before the open; the instrument master downloaded and indexed nightly (it’s too large to parse per lookup); WebSocket 2.0 subscribed in the lightest mode your analytics need — full quote only for instruments where depth actually matters; per-strike OI baselines refreshed from the prior session; and an end-of-day job that reconciles what streamed against candle data, because silent gaps in a binary feed otherwise go unnoticed until an analytic looks wrong a week later. Alert on login failure and on first-tick latency after 9:15 — those two signals catch nearly every real incident.
Angel One API — data FAQ
Is the angel one api free? Yes — no monthly charge for data or orders as of August 2026; only the trading account is required.
Does the feed carry open interest? Yes — derivative ticks include OI across NSE, BSE and MCX, feeding everything from basic OI reads to positioning models.
Can the login really run unattended? Yes — TOTP-based, no browser. Tokens expire daily but re-auth is a scripted POST, which is the property unattended collectors choose it for.
Do I need it to use OIData? OIData runs on your own broker connection as well — the difference is you write nothing: link your account and the ready-made layer runs on data streaming from it. SmartAPI is the foundation for building your own.
How do the rate limits behave? They’re per-endpoint-class and generous for streaming plus a disciplined candle schedule; bulk backfills at the open are the pattern that gets throttled, exactly as on every other Indian broker API. Schedule history pulls off-hours and the limits effectively disappear.
What’s the WebSocket’s failure mode? Silent thinning under bad handling of the 9:15 burst — packets you never knew you missed. The end-of-day reconcile against candles in the checklist above exists precisely to catch this.
Where do people stumble first? The binary tick layout. Field offsets are documented but easy to misread by one byte, and the symptom is plausible-looking wrong numbers rather than an error. Validate your decoder against the broker app’s displayed quote for a handful of liquid instruments before trusting anything downstream — ten minutes that saves days.
Is there historical option OI? As elsewhere, strike-wise option OI history is yours to record from the live feed; the candle endpoint’s OI applies to futures. Long-lookback OI studies favour whichever layer has already been recording continuously.
Verdict
Free, full-coverage, and automatable end to end: the angel one api is the pragmatic pick for unattended data work and MCX-inclusive analysis. Its documentation asks more patience than the polished rivals, and the standard analytics remain a build — or a page you simply open.