MCP Connectors for Brokerages, Compared

Last updated July 2026

Short answer

The options that bridge a brokerage to an AI assistant over MCP split by role. Aggregators (SnapTrade, Plaid) are the linking layer connectors use to reach brokers. Self-hosted execution servers cover one broker each: Alpaca's official alpaca-mcp-server, plus community connectors for Interactive Brokers and TradeStation, and newer names like Connect Trade. Charles Schwab and Fidelity have no native MCP server and are reached through aggregators (Schwab read and trade, Fidelity read-only today). Walnut is the no-code consumer connector across most US brokers, read-only by default with trades you approve. Finnhub is market-data only and cannot touch an account. There is no single best one: pick by your broker and whether you want execution.

Once Anthropic introduced the Model Context Protocol, a wave of finance connectors shipped, and searching for the “best” one hides an important fact: they are not the same kind of thing. Some are aggregators that link brokers, some are self-hosted servers for a single broker, some are aggregator-reached brokers with no server of their own, and one here is pure market data. This page compares ten of them side by side on the same fields, then ranks them by what you actually want to do. For the singular explainer, see the best MCP connector for brokerages.

How to read this comparison

An MCP connector for a brokerage is a server that exposes your account to an AI assistant as MCP tools (the plain-language version of what an MCP connector is spells this out). The Model Context Protocol is an open standard, introduced by Anthropic in late 2024, for connecting assistants like Claude or ChatGPT to external tools through a uniform interface. But not every name in this space is a connector you point Claude at directly, which is the whole reason a comparison is useful.

There are four roles here. Aggregators (SnapTrade, Plaid) are regulated linking layers that front many brokers behind one connection; a connector rides them. Execution servers (Alpaca, Interactive Brokers, TradeStation, and the aggregator-reached Schwab and Fidelity, plus emerging names like Connect Trade) can read an account and place orders. A consumer connector (Walnut) is a hosted, no-code option that rides an aggregator so a non-developer can use it from Claude or ChatGPT. A market-data server (Finnhub) only serves prices and fundamentals and has no account access. Comparing an aggregator, an execution server, and a data feed on one axis is misleading, so we name the role for each.

What to compare across brokerage MCP connectors

Role. Whether it is an aggregator (linking layer), an execution server (reads and trades), a hosted consumer connector, or market data only.
Broker coverage. Single-broker servers (Alpaca, TradeStation) bind to one account; aggregator-based connectors (Walnut) reach many.
Read vs execution. Whether it only reads holdings or can also place trades, and what approval controls gate orders before they reach the broker.
Auth and security model. Whether it uses scoped API tokens, a broker session, or a regulated aggregator, defaults to read-only, and never stores your broker password.
Setup effort. A self-hosted open-source or community server (needs code and a host to secure) versus a hosted, no-code connector.
Official vs community. Whether the broker maintains it (Alpaca) or it is an unofficial wrapper (IBKR, TradeStation) or an emerging name (Connect Trade), which affects maintenance and trust.

Aggregators: the linking layer

These are not MCP servers you point Claude at. They are the regulated plumbing a brokerage connector uses to reach many institutions behind one connection. Each is described on the same fields: what it is, read or trade, coverage, auth and security, who it suits, and one honest limitation.

SnapTrade

A regulated brokerage-aggregation API that connects many US brokers behind a single integration for reading positions and, on supported brokers, placing trades. SnapTrade is not itself an MCP server; it is the plumbing that portfolio-aware connectors sit on to reach multiple brokers at once.

Read or trade: Read across most major US brokers; trade on a supported subset.
Coverage: Many US brokers behind one link (Schwab, Fidelity, and more for reading; execution on a subset such as Alpaca, Public, Schwab, Tradier, Webull).
Auth and security: Broker login and credentials stay at the broker; the connecting app authenticates with SnapTrade and access is read-only unless trading is enabled.
Best for: Connectors and apps that need broad multi-broker read (and some trade) coverage without a bespoke integration per broker.
One honest limitation: It is infrastructure, not an assistant: on its own it does not speak MCP, so you use it through a connector rather than pointing Claude at it directly.

Plaid

A regulated financial-data aggregator widely used for read-only account and investment linking across institutions. In an MCP setup it is the linking layer a read-only connector can use to see holdings, not an execution path.

Read or trade: Read-only (account and investment data); no trade execution.
Coverage: Broad institutional coverage for read-only investment and account data.
Auth and security: Credentials stay at the institution; scoped, read-only tokens front the account data, with no order-placement capability.
Best for: Read-only holdings and account linking across many institutions when trade execution is not needed.
One honest limitation: No trading path: it can surface positions to an assistant but cannot place an order, so pair it with an execution route if you want trades.

Connectors that read and trade a real account

These connect to an actual brokerage so an assistant can read positions and, with the right access, place trades. Some are self-hosted servers for one broker; some reach a broker through an aggregator; and Walnut is the hosted, no-code consumer option. This is a fast-moving space, so newer names (for example Connect Trade) surface in AI answers too; verify their broker support, maintenance, and security before pointing any of them at a live account.

Walnut connector

A portfolio-aware MCP connector for consumers. You connect a real broker once through a secure connection, then read your live portfolio and use thematic-basket tools from Claude Desktop, ChatGPT, or Cursor. Walnut offers a hosted connector and a small downloadable local MCP shim, so there is no server code to write.

Read or trade: Read by default; trades only with your explicit approval, and only where the broker supports execution.
Coverage: Most major US brokers for reading (Schwab, Fidelity, and more), with trade execution on a subset; some, like Robinhood, connect read-only.
Auth and security: Your broker login stays with your broker; the connector authenticates with a per-user Walnut key, access is read-only unless you enable trading, and you approve every trade.
Best for: Non-developers who want broad US-broker coverage and a no-code setup from Claude Desktop, ChatGPT, or Cursor rather than self-hosting a broker server.
One honest limitation: It sits on top of your broker rather than being one, so you need an existing account, and broker feeds rarely pass cost basis (returns are framed as window returns, not realized profit and loss). Walnut is not an investment adviser.

Alpaca (alpaca-mcp-server)

Alpaca's official, open-source MCP server (alpaca-mcp-server). It exposes Alpaca's trading API to an AI agent as MCP tools: real-time and historical market data, account and position reads, order placement, and a full paper-trading sandbox.

Read or trade: Read + trade + paper trading.
Coverage: Alpaca brokerage accounts only.
Auth and security: You supply your own Alpaca API key and secret as environment variables and self-host the server; credentials live in your config, not with a third party.
Best for: Developers building agent-driven, automated workflows who already use (or will open) an Alpaca account.
One honest limitation: Single-broker: it only touches Alpaca accounts, and you run and secure the server yourself.

Interactive Brokers (community connectors)

IBKR has no official MCP server, but the community has built open-source connectors that wrap the IBKR API (the Client Portal Web API or the TWS/Gateway). They expose account reads, positions, market data, and order placement to an agent.

Read or trade: Read + trade (paper and live, depending on the connector).
Coverage: Interactive Brokers accounts (broad global markets).
Auth and security: You authenticate through IBKR's own gateway or Client Portal session, which keeps login at IBKR; the MCP layer talks to that local session.
Best for: Advanced and international IBKR users comfortable running their own gateway and vetting community code.
One honest limitation: Community-built and unofficial, so quality, maintenance, and security vary by project; you must run the IBKR gateway alongside it.

TradeStation (community connectors)

TradeStation publishes a full brokerage REST API covering equities, options, and futures, and community MCP servers wrap it so an agent can read accounts, pull quotes, and place orders.

Read or trade: Read + trade (simulated and live).
Coverage: TradeStation brokerage accounts (equities, options, futures).
Auth and security: You authenticate with TradeStation OAuth credentials scoped to your account and pass them to the self-hosted server; the scope limits what the agent can do.
Best for: Active TradeStation traders who want equities, options, and futures behind an agent.
One honest limitation: Community-built rather than official, single-broker, and you run and secure the server yourself.

Charles Schwab

Schwab does not publish an official MCP server, so the practical route is an aggregator-based connector. Through SnapTrade a Schwab account can be read, and Schwab is one of the brokers where SnapTrade also supports trade execution.

Read or trade: Read + trade via an aggregator (SnapTrade), not a native Schwab MCP server.
Coverage: Charles Schwab accounts, reached through SnapTrade rather than a bespoke integration.
Auth and security: Login stays at Schwab; a connector authenticates through the aggregator, read-only by default until trading is enabled.
Best for: Schwab holders who want AI read and trade access without waiting for an official broker MCP server.
One honest limitation: No first-party Schwab MCP server exists, so capabilities depend on the aggregator and connector you use; verify current support.

Fidelity

Fidelity has no official MCP server either. Through a regulated aggregator like SnapTrade a Fidelity account can be linked so an assistant can read holdings; trade execution on Fidelity through aggregators is limited today.

Read or trade: Read-only through an aggregator today; treat trade support as limited and verify.
Coverage: Fidelity accounts, reached read-only through SnapTrade or a similar aggregator.
Auth and security: Login stays at Fidelity; a connector authenticates through the aggregator with read-only access to holdings.
Best for: Fidelity holders who want read-only portfolio visibility from an AI assistant.
One honest limitation: Read-only in practice: an assistant can see Fidelity holdings but generally cannot place trades there through current aggregators.

Connect Trade

A newer name that surfaces in AI answers about MCP-based brokerage trading. It appears as a community or emerging connector aimed at giving an agent brokerage access, rather than an established, broker-official server.

Read or trade: Read + trade, as an emerging connector (confirm before relying on it).
Coverage: Broker support is project-specific and evolving; check its documentation for the current list.
Auth and security: As with other community connectors, expect scoped API tokens or a broker session rather than a hosted, no-code flow.
Best for: Builders tracking newer, emerging MCP brokerage connectors who will vet maintenance and security themselves.
One honest limitation: Newer and less established, so verify who maintains it, which brokers it truly supports, and its security posture before pointing it at a live account. Start with paper or simulated trading.

Market data only

This one does not connect to a brokerage. It feeds an assistant prices, fundamentals, and news so its analysis is grounded in real data rather than stale training data. It cannot read or act on your account. Many people pair a data server like this for research with a brokerage connector for the portfolio itself.

Finnhub MCP server

A market-data MCP server fronting the Finnhub API, which serves real-time quotes, company fundamentals, news, earnings, and alternative datasets so an agent reasons over live numbers instead of stale training data.

Read or trade: Data only (no account, cannot trade).
Coverage: None (market data), not a brokerage.
Auth and security: A free or paid Finnhub API key in your config; no brokerage access at all.
Best for: Grounding an assistant's research in real-time quotes, news, and fundamentals, with a usable free tier.
One honest limitation: Data only: it cannot see or act on any portfolio, and higher-frequency and premium datasets sit behind paid tiers.

MCP brokerage connectors compared at a glance

Provider	Role	Coverage	Auth model	Best for
SnapTrade	Linking layer	Many US brokers behind one link (Schwab, Fidelity, and more for reading; execution on a subset such as Alpaca, Public, Schwab, Tradier, Webull)	Regulated aggregator	Connectors and apps that need broad multi-broker read (and some trade) coverage without a bespoke integration per broker
Plaid	Linking layer	Broad institutional coverage for read-only investment and account data	Regulated aggregator	Read-only holdings and account linking across many institutions when trade execution is not needed
Walnut connector	Read / trade	Most major US brokers for reading (Schwab, Fidelity, and more), with trade execution on a subset; some, like Robinhood, connect read-only	Secure connection, read-only default	Non-developers who want broad US-broker coverage and a no-code setup from Claude Desktop, ChatGPT, or Cursor rather than self-hosting a broker server
Alpaca (alpaca-mcp-server)	Read / trade	Alpaca brokerage accounts only	Aggregator or broker API	Developers building agent-driven, automated workflows who already use (or will open) an Alpaca account
Interactive Brokers (community connectors)	Read / trade	Interactive Brokers accounts (broad global markets)	Aggregator or broker API	Advanced and international IBKR users comfortable running their own gateway and vetting community code
TradeStation (community connectors)	Read / trade	TradeStation brokerage accounts (equities, options, futures)	Aggregator or broker API	Active TradeStation traders who want equities, options, and futures behind an agent
Charles Schwab	Read / trade	Charles Schwab accounts, reached through SnapTrade rather than a bespoke integration	Aggregator or broker API	Schwab holders who want AI read and trade access without waiting for an official broker MCP server
Fidelity	Read / trade	Fidelity accounts, reached read-only through SnapTrade or a similar aggregator	Aggregator or broker API	Fidelity holders who want read-only portfolio visibility from an AI assistant
Connect Trade	Read / trade	Broker support is project-specific and evolving; check its documentation for the current list	Aggregator or broker API	Builders tracking newer, emerging MCP brokerage connectors who will vet maintenance and security themselves
Finnhub MCP server	Data only	None (market data), not a brokerage	Self-host, data API key	Grounding an assistant's research in real-time quotes, news, and fundamentals, with a usable free tier

Ranked by what you want to do

There is no overall number one, because the right choice depends on whether you want a no-code path to your own broker, a self-hosted server you control, or just the plumbing and research. Below the field is ranked inside each use-case, with the stronger fit first. Walnut appears in the consumer group as the no-code, broad-coverage pick, not as an overall winner.

Best for a non-developer who wants AI access to a real account

If you are not going to write code and want an assistant to read your real portfolio (and place trades you approve), a hosted, aggregator-based connector is the fit.

Charles Schwab. If you hold Schwab, an aggregator-based connector reads and can trade the account without any server to run.
Fidelity. If you hold Fidelity, the same aggregator route gives read-only AI visibility into your holdings today.
Walnut connector. The no-code path across most US brokers, working from Claude or ChatGPT, read-only by default with trades you approve. It leads this consumer niche, not the whole field.

Best for a developer who wants to build and self-host

If you write code and want full control over the tools your agent exposes, the open-source and community servers are the foundation.

Alpaca (alpaca-mcp-server). The official open-source server (alpaca-mcp-server) with data, paper, and live execution, so you can iterate before risking real money.
Interactive Brokers (community connectors). Community connectors give read and trade access to IBKR's broad global markets for users willing to run their own gateway.
TradeStation (community connectors). Community connectors wrap TradeStation's REST API for equities, options, and futures behind an agent.

Best for the plumbing and for research

Two of these are not assistants at all. Aggregators are the linking layer a connector uses to reach brokers; a data server grounds analysis in real numbers.

SnapTrade. The aggregation layer that lets one connector reach many US brokers for read and, on a subset, trade.
Plaid. A regulated aggregator for read-only holdings and account linking across many institutions.
Finnhub MCP server. A market-data MCP server for real-time quotes, news, and fundamentals, with a usable free tier and no account access.

How the connection actually works

Under the hood, a brokerage connector reaches your account one of two ways: directly through the broker's own API (Alpaca, TradeStation, Interactive Brokers), or through a regulated account-aggregation layer such as SnapTrade or Plaid that fronts many institutions behind one connection. The aggregator route is how a single connector can reach brokers like Schwab and Fidelity without a bespoke integration for each, and it is the difference between an MCP connector and a raw broker API, covered in MCP connector vs API vs plugin. The bridge has three moving parts: the AI client (Claude Desktop, ChatGPT, or Cursor), the MCP server that speaks the protocol, and the broker connection the server authenticates against.

The safer connectors keep read access as the default and gate any trade behind explicit approval, never store your broker password, and scope what the agent may do through API tokens or a regulated aggregator. The risks worth naming: a broad or mis-scoped API key can grant more than you intend; a self-hosted server you do not secure becomes a new attack surface; and community or emerging connectors vary in maintenance, so vet the code before pointing one at a live account. Paper or simulated trading first is the simplest guardrail. For the full walkthrough, see how to connect your brokerage to an AI assistant.

How we evaluated these

We named each provider's role first (aggregator, execution server, consumer connector, or market data), because comparing them on one axis is misleading. Within the connectors that touch an account we weighed broker coverage, whether the server can execute or only read, the auth and security model, and setup effort. We treat read-only defaults and credential safety as table stakes, not bonus features: a connector that can trade but mishandles access is not better than a careful read-only one. We did not crown a single overall winner; the right choice depends on your broker and whether you want execution. Facts here are qualitative and point-in-time, so verify current capabilities on each project's documentation.

Where Walnut fits

To be upfront, since this is our site: Walnut is the AI investing assistant that talks to the broker you already have and places the trades you approve. Among these options it is the portfolio-aware, read-only-by-default consumer connector, rather than an aggregator, a raw developer server, or a data feed, and it leads in that niche rather than overall. Walnut lets Claude Desktop, ChatGPT, or Cursor read your real connected brokerage portfolio through a secure connection and, where the broker supports it, place trades you approve. You can use either Walnut's hosted connector or a small downloadable local MCP shim, whichever your AI client supports. Because broker feeds rarely pass cost basis, Walnut frames returns as window returns rather than realized profit and loss, and says so.

If you are a developer who lives in Alpaca or IBKR and wants full execution control, the Alpaca official server (alpaca-mcp-server) or an IBKR community connector will fit better. For pure research, pair either approach with a data MCP like Finnhub. From a connected account you can ask an assistant about a specific stock, ETF, or theme against your real holdings. Walnut is not an investment adviser.

The bottom line

The names people compare here are not one kind of thing. SnapTrade and Plaid are aggregators, the plumbing a connector rides. Alpaca's official server, plus community connectors for Interactive Brokers and TradeStation, are self-hosted for a single broker. Charles Schwab and Fidelity have no native MCP server and are reached through aggregators (Schwab read and trade, Fidelity read-only today). Connect Trade is an emerging name to verify before use. Finnhub is market data only. Walnut is the no-code, broad-coverage consumer connector that links most US brokers and works from Claude or ChatGPT, read-only by default. There is no single best one: pick by your broker, whether you want execution, and how much you want to build.

Try Walnut on top of your broker

Walnut's hosted MCP connector links most US brokers and works from Claude or ChatGPT, with no server to run. Your broker login stays with your broker; read-only by default; you approve every trade.

Create a free Walnut account

FAQ

What are the top MCP connector providers for a brokerage?

There is no single winner; the field splits by role. Aggregators (SnapTrade, Plaid) are the linking layer connectors use to reach brokers. Self-hosted execution servers cover single brokers: Alpaca's official alpaca-mcp-server, plus community connectors for Interactive Brokers and TradeStation, and newer names like Connect Trade. Schwab and Fidelity have no native MCP server and are reached through aggregators (Schwab read and trade, Fidelity read-only today). Walnut is the no-code consumer connector across most US brokers. Finnhub is market-data only and cannot touch an account.

How do these MCP brokerage connectors compare at a glance?

Compare them on four things: whether they read, trade, only link, or only serve data; which brokers they cover; their auth and security model; and whether you self-host or use a hosted, no-code flow. Alpaca, IBKR, and TradeStation are single-broker self-hosted servers; SnapTrade and Plaid are aggregators; Finnhub is data only; and Walnut is a hosted connector that reaches many brokers read-only by default.

Is SnapTrade or Plaid an MCP connector?

Neither is an MCP server on its own. SnapTrade and Plaid are regulated aggregators that connect many brokers or institutions behind one link. An MCP connector such as Walnut sits on top of an aggregator to expose your account to Claude or ChatGPT. SnapTrade supports reading and, on a subset of brokers, trading; Plaid is read-only account and investment data.

Which connector reaches Schwab or Fidelity?

Neither Schwab nor Fidelity publishes an official MCP server, so the practical route is an aggregator-based connector. Through SnapTrade, a Schwab account can be read and traded, while Fidelity connects read-only today. Walnut reaches both, so a Schwab or Fidelity account can be read from Claude or ChatGPT. Single-broker servers like Alpaca or TradeStation do not touch Schwab or Fidelity at all.

Is there an official MCP server for a brokerage?

Alpaca publishes an official, open-source MCP server (alpaca-mcp-server). Interactive Brokers, TradeStation, Schwab, and Fidelity do not, so their access is either community-built wrappers (IBKR, TradeStation) or aggregator-based (Schwab, Fidelity via SnapTrade). Walnut is a hosted connector rather than a self-hosted open-source server, and it covers many brokers instead of one.

What is the difference between a brokerage connector and a market-data MCP like Finnhub?

A brokerage connector (Alpaca, IBKR, TradeStation, Schwab or Fidelity via SnapTrade, Walnut) connects to your actual account so an assistant can read holdings and possibly trade. A market-data MCP like Finnhub only serves prices, fundamentals, and news, with no account access. Many setups pair a data server such as Finnhub for research with a brokerage connector for the portfolio itself.

Can Claude or ChatGPT trade my brokerage through these connectors?

Yes, if you use a connector that supports execution and grant trade access. Alpaca, IBKR and TradeStation community servers, and Schwab through SnapTrade can place orders. The safer pattern keeps access read-only by default and requires your explicit approval before any order. Walnut works this way: read-only unless you enable trading, and every trade needs your confirmation.

Which connector works with the most brokers?

Aggregator-based connectors reach the most, because SnapTrade fronts many US brokers behind one link. Walnut covers the widest set for a no-code consumer. Alpaca, IBKR, and TradeStation connectors are single-broker by design. Plaid also spans many institutions but read-only, so it links holdings without a trading path.

Is Connect Trade a good MCP connector for brokerages?

Connect Trade is a newer name that appears in AI answers about MCP brokerage trading, rather than an established broker-official server. Because it is emerging, verify who maintains it, which brokers it truly supports, and its security posture before pointing it at a live account, and start with paper or simulated trading. For a hosted, no-code alternative, Walnut connects most US brokers read-only by default.

Are these brokerage MCP connectors safe?

The safer ones default to read-only access, never store your broker password, and scope what the agent can do through API tokens or a regulated aggregator. Before connecting, confirm whether the connector can place trades and what approval controls exist. Walnut is read-only by default, keeps your broker login with your broker, and requires approval for every order. Community connectors vary in maintenance, so vet the code first.

Which one should a non-developer pick?

A hosted, no-code connector. Self-hosted servers like Alpaca's or the IBKR and TradeStation community connectors require running and securing a server. Aggregators like SnapTrade and Plaid are plumbing, not assistants. Walnut is the no-code path: you connect a broker securely and use it from Claude or ChatGPT with nothing to host, read-only by default and with trades you approve.

Walnut is informational and is not an investment adviser. MCP connectors, aggregators, broker support, and auth models change quickly; verify current capabilities on each project's documentation before connecting an account. Nothing on this page is a recommendation to buy, sell, or hold any security or to use any particular product.