Alpaca Review 2026: API-First Broker for Coders

What Alpaca actually is

Alpaca is a US-regulated broker-dealer built for people who write code, not people who tap buttons. The entity that holds your securities is Alpaca Securities LLC, a FINRA member and SIPC participant, and crypto runs through a separate entity, Alpaca Crypto LLC. That split matters: equities and crypto sit under different regulatory umbrellas, with different protections and different disclosures. If you want the FINRA BrokerCheck record or the SIPC coverage page, both are public and worth reading before you fund anything.

The company runs two distinct products. The first is the Trading API, which is what this review covers: a self-directed brokerage account you drive through REST and WebSocket endpoints. The second is the Broker API, a white-label product that lets other companies embed brokerage into their own apps. Those are not the same thing, and a lot of confusion online comes from people reading Broker API marketing and assuming it describes the retail trading account. It does not.

The headline pitch is commission-free US equities and ETFs, with crypto priced on a spread rather than a per-trade commission. The site lists a starting price of $0/mo, and the company says it has more than 3 million accounts. What Alpaca is really selling is programmatic access: order placement, streaming market data, and account management over HTTP and WebSocket, with SDKs in Python, JavaScript, Go, and C#. If you have ever wanted to run a strategy on a schedule without clicking through a phone app, that is the entire value proposition.

What it is not: a consumer investing app. There is no robo-advisor, no curated portfolio, no goal tracker. You bring the logic. Alpaca brings the plumbing.

Account tiers and pricing in 2026

Alpaca splits accounts into individual and business types. The individual account is the one most readers will open. The provider states there is no live account minimum for individual accounts, which is unusual among API-first brokers and lowers the barrier for someone who wants to test a strategy with a small balance. Business accounts are aimed at entities and come with their own onboarding requirements. Check the current account documentation for the exact paperwork, because entity onboarding is where requirements tend to change.

Market data is where the pricing structure actually gets interesting, and it is the part people miss. The free tier gives you IEX-only data. IEX is a single exchange, so the quote you see is not the consolidated tape. For a strategy that only needs approximate prices, that is fine. For anything sensitive to the national best bid and offer, IEX-only is a real limitation, and it is the single most common source of confusion for new users who wonder why their fills look different from what a mainstream broker shows.

The paid data tier, Algo Trader Plus, provides full SIP data, meaning the consolidated feed across US exchanges. Alpaca does not publish a single universal price for that tier in a way that applies to every account type and promotion, so I would not quote a number here. Check the provider's current pricing page for the exact monthly figure and whether it is waived based on trading activity or account type.

Margin is available on eligible accounts, with rates that vary by tier and by how much you borrow. Short borrowing carries its own fee schedule, and hard-to-borrow names can be expensive. Crypto is spread-based: instead of a commission line item, the cost is embedded in the bid-ask spread you trade against. That is a legitimate model, but it means the true cost of a crypto trade is not visible as a fee on your statement. To compare crypto costs across venues, you have to measure the spread, not the commission.

Because the verified fee data for this review does not include specific figures, I am deliberately not printing dollar amounts for margin, short borrow, or crypto. Name the fee type, then go read the schedule. Any review that quotes a margin rate without a date stamp is probably stale.

The API: what works and what bites

The core surface is REST for request-response operations and WebSocket for streaming. REST handles account state, order submission, position queries, and historical bars. WebSocket handles real-time trades, quotes, and order updates. The SDKs wrap both, and they are reasonably thin, which is good: thin wrappers mean fewer surprises when the underlying API changes.

Authentication uses API keys with scoping. You can issue keys limited to market data only, or trading, or both. That is a meaningful security feature and one of the better design decisions in the product. If you are running a data-collection script on a server, you can give it a read-only key and never expose trading permissions. Use it. A key with full trading scope sitting in a config file on a VPS is a bad day waiting to happen.

Paper trading is the feature that sells the platform to most developers, and it deserves an honest assessment. The paper environment mirrors the live API surface closely: same endpoints, same order types, same response shapes. That parity is the point, and it means code written against paper generally runs against live with minimal changes. Where paper diverges is in fills. Paper fills are simulated against the market data feed, and simulation is not matching. In thin markets, during fast moves, or on orders that would queue behind other participants in real life, paper fills can be more optimistic than what you would actually get. Treat paper as a correctness test for your code, not as a performance forecast for your strategy. Anyone who has run a backtest that looked great and a live account that did not already knows this, but it is worth stating plainly.

Rate limits are published in the docs, and they differ by endpoint and by data tier. The free data tier has tighter limits than the paid tier. If you are polling REST endpoints in a loop, you will hit them. The fix is to use WebSocket streams for anything real-time and reserve REST for state changes. The docs list the specific limits per endpoint, and they are worth reading before you architect anything.

Error handling is where a lot of first integrations stumble. The API returns structured error codes, and the docs enumerate them. The ones that bite in practice are the order rejection codes: insufficient buying power, symbol not tradable, order would violate a pattern day trader restriction, and market closed. Your code needs to handle these as normal outcomes, not exceptions. A rejected order is not a bug in the API; it is the API telling you something about your account or your order. Log the code, not just the message.

WebSocket reconnect behavior is the other thing to plan for. Connections drop. Networks blip. Servers restart. The client libraries handle some reconnection, but you should assume you will need to re-authenticate and re-subscribe after a drop, and you should have a gap-detection strategy: if you miss a window of streaming data, you need to backfill from REST historical endpoints. Alpaca publishes the stream endpoints and subscription format in the docs. Build the reconnect path before you need it, because you will need it.

Order types, routing, and execution quality

The supported order set is broad for a retail API broker. Market, limit, stop, stop-limit, trailing stop, bracket orders, and OCO (one-cancels-other) are all available. Fractional and notional orders are supported, which matters for anyone running a dollar-based allocation strategy rather than a share-based one. Extended-hours trading is available on eligible orders, though liquidity in extended sessions is thin and spreads widen, so the execution you get outside regular hours is not the execution you get inside them.

Bracket and OCO orders are the ones that make Alpaca genuinely useful for systematic trading. A bracket order attaches a take-profit and a stop-loss to an entry in one submission, and the platform manages the exits. Doing that manually across a REST API is error-prone; having it as a native order type removes a whole class of bugs from your code. If your strategy has defined exits, use brackets.

Pattern day trader handling is worth understanding before you start. The PDT rule applies to margin accounts with less than the regulatory threshold and limits you to a set number of day trades in a rolling window. Alpaca enforces it, and the API will reject orders that would breach it. If you are running a high-frequency strategy in a small account, this is the constraint that will stop you, not the API. Cash accounts have different rules and different settlement timing. Read the account documentation for your specific account type.

On routing and execution quality: Alpaca routes US equity orders to market centers and, per its disclosures, may receive payment for order flow. That is standard practice across the retail brokerage industry, and it is disclosed. What it means for you is that the price you get is the price the market center provides, and Alpaca's disclosures describe how routing decisions are made and what price improvement statistics look like. If execution quality is central to your strategy, read those disclosures rather than assuming. The company publishes the relevant documents; they are not hidden.

One honest limitation: Alpaca is not a smart order router in the institutional sense, and it does not offer the venue selection controls that a professional trading system would. You get the routing the broker provides. For most retail algorithmic strategies that is fine. For anything where venue choice is part of the edge, it is a constraint to weigh.

Alpaca vs Interactive Brokers vs M1

These three serve different people, and comparing them on a single axis misses the point.

Interactive Brokers is the broadest of the three. It covers far more asset classes and markets, has a long track record, and offers an API that institutional and semi-professional traders use. Its pricing structure is more complex, with commissions on many products and a tiered model. If you need global markets, futures, or options depth, IBKR is the more capable platform. The tradeoff is complexity: the API is powerful but heavier, and the account structure has more moving parts. For a hobbyist running a US equity strategy, IBKR is often more platform than the job requires.

M1 Finance is the opposite end. It is a consumer investing app built around automated portfolio allocation, pies, and a long-term holding philosophy. It has no meaningful public trading API for algorithmic order flow. If your goal is to set an allocation and let it run, M1 is a clean product. If your goal is to submit orders programmatically, it is the wrong tool. Our M1 Finance tool page covers what it does well.

Alpaca sits between them. It is narrower than IBKR on asset coverage and market access, and it is far more programmable than M1. Its advantage is developer experience: clean REST and WebSocket APIs, thin SDKs, a paper environment that mirrors live, and no account minimum for individual accounts. Its disadvantage is breadth: if you need options strategies, futures, or international exchanges, you will hit the edges of what Alpaca offers. For a comparison of two more consumer-facing brokers, see our Interactive Brokers vs Webull breakdown.

On cost per trade, all three are competitive for US equities in different ways, and the details change often enough that any number printed here would age badly. Check each provider's current fee schedule. On API quality, Alpaca is the most developer-friendly of the three for straightforward US equity and crypto strategies. On asset coverage, IBKR wins. On simplicity for buy-and-hold, M1 wins.

Who it's for, and who should skip it

Alpaca fits a specific person: someone who writes code and wants to trade US equities, ETFs, or crypto through that code. That includes algorithmic traders testing strategies, fintech builders prototyping products, and quant hobbyists who want a real brokerage behind their scripts rather than a simulated one. The paper environment makes it easy to start without funding, and the lack of an individual account minimum means you can go live small.

It also fits people who want to automate a simple rule-based strategy: a rebalancing script, a dollar-cost-averaging bot, a signal-driven entry with bracket exits. You do not need to be a professional quant. You need to be comfortable with an API key, a JSON response, and an error log.

Who should skip it: buy-and-hold investors who want a polished mobile app, curated portfolios, and a support line staffed for consumer questions. Alpaca's interface is functional, not delightful, and its support model is built around developers. If you have never opened a terminal, the platform will feel like work. That is not a flaw; it is the design. But it means the product is wrong for a large share of retail investors, and pretending otherwise would be dishonest.

Also skip it if your strategy depends on options, futures, or non-US markets. Those are outside what the Trading API covers today, and no amount of clever code will route around an unsupported asset class.

Verdict

Alpaca does one thing well and does not pretend to do the rest. As an API-first brokerage, it offers clean REST and WebSocket access, SDKs in the languages developers actually use, scoped API keys, a paper environment that mirrors live endpoints, and a broad order-type set including brackets and OCO. The constraints are real: IEX-only data on the free tier, simulated paper fills that can flatter a strategy, rate limits that punish naive polling, and no options, futures, or international markets. The fee schedule and disclosures are the place to check current costs, because the specifics move.

For a developer who wants to trade US equities and crypto programmatically, it is the most direct path from code to a live order. For everyone else, a conventional broker is the better fit.

How this review was researched

This review is based on Alpaca's own published materials: the official site at alpaca.markets, the API documentation at docs.alpaca.markets (including the rate limit tables and order type matrix), the provider's fee schedule and disclosures, the FINRA BrokerCheck record for Alpaca Securities LLC, and the SIPC coverage page. No account was opened, no funds were deposited, and no orders were placed in preparing this review. Where a fee or limit is not stated in the sources above, this review names the fee type without a figure and directs readers to the provider's current schedule. Our methodology page explains how we assess products, and our disclosure page covers how this publication is funded. For a broader look at the category, see our best stock apps of 2026 roundup.