LangAlpha

LangAlpha is an open-source AI agent harness for financial research and investment decision-making, built on LangGraph and LangChain. It brings the persistent-workspace paradigm from software engineering to investing: each research goal gets its own sandbox where the agent accumulates findings across sessions, threads, and analyses. The system supports programmatic tool calling (PTC), where the agent writes and executes Python code inside cloud sandboxes to process financial data rather than dumping raw data into the LLM context window.

• Persistent Workspaces — Each workspace maps to a Daytona cloud sandbox with structured directories and an agent.md memory file that compounds research across sessions and threads.
• Programmatic Tool Calling (PTC) — The agent writes and executes Python to process financial data from MCP servers, enabling complex multi-step analysis while dramatically reducing token waste.
• Financial Data Ecosystem — Multi-tier provider hierarchy with native tools for quick lookups (company overviews, SEC filings, market indices) and MCP servers for bulk data processing, charting, and multi-year analysis.
• 23 Pre-built Financial Research Skills — Activatable by slash command or auto-detection, covering DCF models, initiating coverage reports, earnings analysis, morning notes, and document generation (PDF, DOCX, PPTX, XLSX).
• Agent Swarm — Parallel async subagents with isolated context windows, mid-execution steering, checkpoint-based resume, and live progress monitoring in the UI.
• Live Steering — Send follow-up messages while the agent is working to course-correct, clarify, or redirect without waiting for it to finish.
• Multi-Provider Model Layer — Provider-agnostic LLM abstraction supporting Gemini, OpenAI, Anthropic, DeepSeek, and more, with automatic failover and BYOK support.
• Automations — Schedule recurring or one-shot tasks via cron, or set price-triggered automations that fire when a stock or index hits a real-time price condition.
• Finance Research Workbench — Web UI with inline financial charts, TradingView charting, multi-format file viewer, real-time WebSocket market data, shareable conversations, and subagent monitoring.
• Channel Integrations — Use LangAlpha from Slack, Discord, Feishu, and Telegram with full feature support including file upload/download, human-in-the-loop interrupts, and slash commands.
• Security & Workspace Vault — Encryption at rest via pgcrypto, automatic credential leak detection and redaction, sandboxed code execution, and per-workspace secret storage.
• Middleware Stack — 24 composable layers handling skill loading, plan mode, multimodal input, auto-summarization, and context management for long-running agent sessions.

To get started, clone the repository, run make config to configure your LLM and data sources, then make up to launch the full stack with Docker.