Agent-first financial infrastructure, built on Alpaca
Build, operate, and automate market workflows with Alpaca’s AI-ready infrastructure. Alpaca’s MCP servers and CLI give traders, developers, and Broker API partners new ways to connect agents to trading, market data, brokerage operations, and approved actions.
Numbers we’ll back up.
1.5ms
Order processing with OMS v2
99.99%
Uptime since January 2026
$0
Borrow fees on 5,000+ ETB stocks
600k+
Daily orders self-cleared (Q1 2026)
What makes agents on Alpaca powerful
AI-native market access
Bring market data, account context, and trading actions into AI interfaces where users already work.
Brokerage operations with agents
Give teams a faster way to inspect accounts, transactions, and correspondent data through Broker MCP.
One Alpaca infrastructure layer
Connect trading, market data, account management, and brokerage operations through Alpaca’s API-first platform.
Built for testing and scale
Start in paper, Sandbox, or staging environments before moving workflows into production.
Three ways to build with agents
Trading MCP Server
Connect AI tools like Claude, Cursor, ChatGPT, VS Code, Gemini CLI, and other MCP-compatible clients to Alpaca’s Trading API.
Ask questions, analyze markets, monitor portfolios, and place trades across stocks, ETFs, options, and crypto.
Broker MCP Server
Alpaca’s Broker MCP Server helps partners and teams interact with Broker API correspondents through an AI interface.
Agents can query data, inspect accounts, review transactions, and, in Sandbox or staging environments, help build and test operational workflows.
Alpaca CLI
Alpaca’s CLI turns the Trading API into explicit terminal commands with structured JSON output. Use it to submit orders, pull market data, check accounts, manage positions, export data, and automate workflows from scripts, CI pipelines, cron jobs, or AI agent sessions.
Built for control
For trading
- Trading MCP supports toolset filtering to limit what agents can access
- CLI commands are explicit, reproducible, and easy to log or replay.
- Paper trading lets users test before going live.
For building
- Broker MCP requires correspondent access and Client Secret API credentials
- Broker MCP live environments expose a read-only surface
- Sandbox and staging environments support fuller testing workflows.
Ease of use
- Plugins make using any of Alpaca’s MCP servers easy.
- Open Source Skills Library to help your agents run repeatable workflows.
Build with the AI native brokerage infrastructure company
Connect AI agents, terminal automation, and brokerage infrastructure through Alpaca’s MCP servers and CLI