---

TradingAgents：多智能体 LLM 金融交易框架

• [2026-07] TradingAgents v0.3.1 发布，带来正确性与稳定性修复：Alpha Vantage 前视偏差过滤、graph-router 崩溃安全、图形状感知的检查点续跑、可用的加密情绪源、可配置的 LLM 重试预算、Bedrock API-key 认证，以及对 Claude Sonnet 5 / Fable 5 的支持。完整列表见 CHANGELOG.md。
• [2026-06] TradingAgents v0.3.0 发布，包含已验证的数据访问契约、扩展的提供商注册表（NVIDIA、Kimi、Groq、Mistral、Bedrock 以及任何 OpenAI 兼容端点）、FRED 和 Polymarket 数据供应商、当前最新的模型目录以及 CI 门禁。
• [2026-05] TradingAgents v0.2.5 发布，引入基于事实的情绪分析师、覆盖 GPT-5.5 等模型、Qwen/GLM/MiniMax 双区域支持、TRADINGAGENTS_* 环境变量配置与 API-key 自动检测、远程 Ollama 支持、非美股基准测试以及股票代码路径遍历加固。
• [2026-04] TradingAgents v0.2.4 发布，新增结构化输出智能体（研究经理、交易员、投资组合经理）、LangGraph 检查点续跑、持久决策日志、DeepSeek/Qwen/GLM/Azure 提供商支持、Docker 以及 Windows UTF-8 编码修复。
• [2026-03] TradingAgents v0.2.3 发布，支持多语言、GPT-5.4 系列模型、统一模型目录、回测日期保真度及代理支持。
• [2026-03] TradingAgents v0.2.2 发布，覆盖 GPT-5.4/Gemini 3.1/Claude 4.6 模型、五级评分体系、OpenAI Responses API、Anthropic effort 控制以及跨平台稳定性。
• [2026-02] TradingAgents v0.2.0 发布，支持多提供商 LLM（GPT-5.x、Gemini 3.x、Claude 4.x、Grok 4.x）并改进系统架构。
• [2026-01] Trading-R1 技术报告 发布，对应的 终端 预计即将面世。

🎉 TradingAgents 正式发布！我们收到了大量关于本工作的咨询，衷心感谢社区的热情。

因此，我们决定将框架完全开源。期待与您一起构建有影响力的项目！

TradingAgents 是一个多智能体交易框架，模拟真实世界交易公司的运作模式。通过部署由 LLM 驱动的专业智能体——从基本面分析师、情绪专家、技术分析师，到交易员、风险管理团队——该平台协同评估市场状况并形成交易决策。此外，这些智能体还进行动态讨论，以确定最优策略。

TradingAgents 框架专为研究目的而设计。交易表现可能因多种因素而异，包括所选用的底层语言模型、模型温度、交易时段、数据质量以及其他非确定性因素。本框架无意作为金融、投资或交易建议。

我们的框架将复杂的交易任务分解为专业角色。

• 基本面分析师：评估公司财务与绩效指标，识别内在价值与潜在风险信号。
• 情绪分析师：将新闻标题、StockTwits 和 Reddit 讨论聚合为单一情绪读数，以衡量短期市场情绪。
• 新闻分析师：监控全球新闻和宏观经济指标，解读事件对市场状况的影响。
• 技术分析师：利用技术指标（如 MACD 和 RSI）检测交易模式并预测价格走势。

• 包含看多与看空两位研究员，他们对分析师团队提供的见解进行批判性评估。通过结构化辩论，平衡潜在收益与内在风险。

• 汇总分析师和研究员报告，做出明智的交易决策，确定交易的时机与规模。

风险管理与投资组合经理

• 通过评估市场波动性、流动性及其他风险因子，持续评估投资组合风险。风险管理团队评估并调整交易策略，向投资组合经理提交评估报告以做最终决策。
• 投资组合经理批准/拒绝交易提案。若批准，订单将发送至模拟交易所并执行。

克隆 TradingAgents：

git clone https://github.com/TauricResearch/TradingAgents.git cd TradingAgents

使用您喜欢的环境管理器创建虚拟环境：

conda create -n tradingagents python=3.12 conda activate tradingagents

安装包及其依赖：

或者，使用 Docker 运行：

cp .env.example .env # 添加您的 API keys docker compose run --rm tradingagents

对于使用 Ollama 的本地模型：

docker compose --profile ollama run --rm tradingagents-ollama

TradingAgents 支持多种 LLM 提供商。为您选择的提供商设置 API key：

export OPENAI_API_KEY=... # OpenAI (GPT) export GOOGLE_API_KEY=... # Google (Gemini) export ANTHROPIC_API_KEY=... # Anthropic (Claude) export XAI_API_KEY=... # xAI (Grok) export DEEPSEEK_API_KEY=... # DeepSeek export DASHSCOPE_API_KEY=... # Qwen — 国际版 (dashscope-intl.aliyuncs.com) export DASHSCOPE_CN_API_KEY=... # Qwen — 中国版 (dashscope.aliyuncs.com) export ZHIPU_API_KEY=... # GLM via Z.AI (国际版) export ZHIPU_CN_API_KEY=... # GLM via BigModel (中国区, open.bigmodel.cn) export MINIMAX_API_KEY=... # MiniMax — 全球 (api.minimax.io) export MINIMAX_CN_API_KEY=... # MiniMax — 中国区 (api.minimaxi.com) export OPENROUTER_API_KEY=... # OpenRouter export ALPHA_VANTAGE_API_KEY=... # Alpha Vantage

对于 Azure OpenAI，请将 .env.enterprise.example 复制为 .env.enterprise 并填入您的凭据。

对于 AWS Bedrock，请使用 pip install ".[bedrock]" 安装额外依赖，将 llm_provider 设为 "bedrock"，配置 AWS 凭据（环境变量、~/.aws/credentials 或 IAM 角色）和 AWS_DEFAULT_REGION，并使用 Bedrock 模型 ID，例如 us.anthropic.claude-opus-4-8-v1:0。

对于本地模型，请使用 llm_provider: "ollama" 配置 Ollama。默认端点为 http://localhost:11434/v1；可通过 OLLAMA_BASE_URL 指向远程 ollama-serve。使用 ollama pull <name> 拉取模型，并在 CLI 中为未默认列出的任何模型选择 “Custom model ID”。

对于任何其他 OpenAI 兼容服务器（vLLM、LM Studio、llama.cpp 或自定义中继），请使用 llm_provider: "openai_compatible"，并通过 backend_url（或 TRADINGAGENTS_LLM_BACKEND_URL）设置端点，例如 vLLM 使用 http://localhost:8000/v1，LM Studio 使用 http://localhost:1234/v1。模型由您的服务器提供。本地服务器无需 key；若端点需要，请设置 OPENAI_COMPATIBLE_API_KEY。

或者，将 .env.example 复制为 .env 并填入您的 keys：

启动交互式 CLI：

tradingagents # 安装的命令 python -m cli.main # 替代方式：直接从源码运行

您将看到一个屏幕，可在其中选择所需的股票代码、分析日期、LLM 提供商、研究深度等。

TradingAgents 适用于 Yahoo Finance 覆盖的任何市场，使用带有交易所后缀的股票代码。公司身份和 alpha 基准会自动按市场解析。

• 美国：AAPL、SPY
• 香港：0700.HK · 东京：7203.T · 伦敦：AZN.L
• 印度：RELIANCE.NS、.BO · 加拿大：.TO · 澳大利亚：.AX
• 中国 A 股：上海 .SS，深圳 .SZ（例如贵州茅台 600519.SS）
• 加密货币：BTC-USD、ETH-USD

将出现一个界面，随着结果加载显示，让您跟踪智能体执行进度。

我们基于 LangGraph 构建 TradingAgents，以确保灵活性和模块化。框架支持多种 LLM 提供商：OpenAI、Google、Anthropic、xAI、DeepSeek、Qwen（阿里云 DashScope，国际和国内端点）、GLM（智谱）、MiniMax（全球 + 国内）、OpenRouter、用于本地模型的 Ollama，以及用于企业的 Azure OpenAI。

要在代码中使用 TradingAgents，可导入 tradingagents 模块并初始化 TradingAgentsGraph() 对象。.propagate() 函数将返回一个决策。可运行 main.py，以下是一个快速示例：

from tradingagents.graph.trading_graph import TradingAgentsGraph from tradingagents.default_config import DEFAULT_CONFIG

ta = TradingAgentsGraph(debug=True, config=DEFAULT_CONFIG.copy())

前向传播

_, decision = ta.propagate("NVDA", "2026-01-15") print(decision)

您还可以调整默认配置，以设置自己的 LLM 选择、辩论轮次等。

from tradingagents.graph.trading_graph import TradingAgentsGraph from tradingagents.default_config import DEFAULT_CONFIG

config = DEFAULT_CONFIG.copy() config["llm_provider"] = "openai" # 例如 openai, google, anthropic, deepseek, groq, ollama；openai_compatible 涵盖任何 OpenAI 兼容端点 (vLLM, LM Studio, llama.cpp, ...) config["deep_think_llm"] = "gpt-5.5" # 用于复杂推理的模型 config["quick_think_llm"] = "gpt-5.4-mini" # 用于快速任务的模型 config["max_debate_rounds"] = 2

ta = TradingAgentsGraph(debug=True, config=config) _, decision = ta.propagate("NVDA", "2026-01-15") print(decision)

所有配置选项见 tradingagents/default_config.py。

TradingAgents 在每次运行时持久化两类状态。

决策日志始终开启。每次完成的运行将其决策追加到 ~/.tradingagents/memory/trading_memory.md。下一次对同一股票代码运行时，TradingAgents 会获取已实现收益率（原始值与相对于 SPY 的 alpha），生成一段反思，并将最近同一股票代码的决策以及近期跨股票代码的经验注入投资组合经理提示，使每次分析都延续此前有效或无效的经验。

可通过 TRADINGAGENTS_MEMORY_LOG_PATH 覆盖路径。

检查点续跑通过 --checkpoint 选择加入。启用后，LangGraph 会在每个节点后保存状态，这样崩溃或中断的运行将从最后成功步骤恢复，而无需重新开始。恢复运行时，您将在日志中看到 Resuming from step N for <TICKER> on <date>；全新运行时则显示 Starting fresh。检查点在成功完成后自动清除。

每个股票代码的 SQLite 数据库位于 ~/.tradingagents/cache/checkpoints/<TICKER>.db（可通过 TRADINGAGENTS_CACHE_DIR 覆盖根目录）。使用 --clear-checkpoints 在运行前重置所有检查点。

tradingagents analyze --checkpoint # 本次运行启用 tradingagents analyze --clear-checkpoints # 运行前重置

config = DEFAULT_CONFIG.copy() config["checkpoint_enabled"] = True ta = TradingAgentsGraph(config=config) _, decision = ta.propagate("NVDA", "2026-01-15")

TradingAgents 是 LLM 驱动的，因此同一股票代码和日期的两次运行结果可能不同。这对于基于语言模型构建的研究工具是正常现象，而非缺陷。差异来自几个不同来源，区分这些来源很有帮助。

语言模型采样是非确定性的。即使在固定温度下，提供商也不保证跨调用输出逐字节相同，而推理模型（默认的 GPT-5.x 系列以及任何思考模式模型）的内部推理本身也是采样产生的，因此差异最大。

实时数据会变动。新闻、StockTwits 和 Reddit 随时间返回不同内容，因此即使对同一历史交易日期，今天运行所见输入也不同于上周运行。固定分析日期可以保持价格和指标窗口不变，但社交媒体和新闻源仍反映“当前”内容。

要减少变异，可降低采样温度。在配置中设置 temperature（或在 .env 中设置 TRADINGAGENTS_TEMPERATURE）；较低的值可使遵循该参数的模型更可重复。当前精选的模型以推理优先，普遍忽略温度参数，因此若需更好的可重复性，可使用非推理模型，并将 Custom model ID 设置为非推理模型。

---

TradingAgents: Multi-Agents LLM Financial Trading Framework

• [2026-07] TradingAgents v0.3.1 released with correctness and stability fixes: Alpha Vantage look-ahead filtering, graph-router crash-safety, graph-shape-aware checkpoint resume, working crypto sentiment sources, a configurable LLM retry budget, Bedrock API-key auth, and Claude Sonnet 5 / Fable 5 support. See CHANGELOG.md for the full list.
• [2026-06] TradingAgents v0.3.0 released with a verified data-access contract, an expanded provider registry (NVIDIA, Kimi, Groq, Mistral, Bedrock, and any OpenAI-compatible endpoint), FRED and Polymarket data vendors, a current-generation model catalog, and a CI gate.
• [2026-05] TradingAgents v0.2.5 released with the grounded Sentiment Analyst, GPT-5.5 etc. model coverage, Qwen/GLM/MiniMax dual-region support, TRADINGAGENTS_* env-var configurability with API-key auto-detection, remote Ollama support, non-US alpha benchmarks, and ticker path-traversal hardening.
• [2026-04] TradingAgents v0.2.4 released with structured-output agents (Research Manager, Trader, Portfolio Manager), LangGraph checkpoint resume, persistent decision log, DeepSeek/Qwen/GLM/Azure provider support, Docker, and a Windows UTF-8 encoding fix.
• [2026-03] TradingAgents v0.2.3 released with multi-language support, GPT-5.4 family models, unified model catalog, backtesting date fidelity, and proxy support.
• [2026-03] TradingAgents v0.2.2 released with GPT-5.4/Gemini 3.1/Claude 4.6 model coverage, five-tier rating scale, OpenAI Responses API, Anthropic effort control, and cross-platform stability.
• [2026-02] TradingAgents v0.2.0 released with multi-provider LLM support (GPT-5.x, Gemini 3.x, Claude 4.x, Grok 4.x) and improved system architecture.
• [2026-01] Trading-R1 Technical Report released, with Terminal expected to land soon.

🎉 TradingAgents officially released! We have received numerous inquiries about the work, and we would like to express our thanks for the enthusiasm in our community.

So we decided to fully open-source the framework. Looking forward to building impactful projects with you!

TradingAgents is a multi-agent trading framework that mirrors the dynamics of real-world trading firms. By deploying specialized LLM-powered agents: from fundamental analysts, sentiment experts, and technical analysts, to trader, risk management team, the platform collaboratively evaluates market conditions and informs trading decisions. Moreover, these agents engage in dynamic discussions to pinpoint the optimal strategy.

TradingAgents framework is designed for research purposes. Trading performance may vary based on many factors, including the chosen backbone language models, model temperature, trading periods, the quality of data, and other non-deterministic factors. It is not intended as financial, investment, or trading advice.

Our framework decomposes complex trading tasks into specialized roles.

• Fundamentals Analyst: Evaluates company financials and performance metrics, identifying intrinsic values and potential red flags.
• Sentiment Analyst: Aggregates news headlines, StockTwits, and Reddit chatter into a single sentiment read to gauge short-term market mood.
• News Analyst: Monitors global news and macroeconomic indicators, interpreting the impact of events on market conditions.
• Technical Analyst: Utilizes technical indicators (like MACD and RSI) to detect trading patterns and forecast price movements.

• Comprises both bullish and bearish researchers who critically assess the insights provided by the Analyst Team. Through structured debates, they balance potential gains against inherent risks.

• Composes reports from the analysts and researchers to make informed trading decisions, determining the timing and magnitude of trades.

Risk Management and Portfolio Manager

• Continuously evaluates portfolio risk by assessing market volatility, liquidity, and other risk factors. The risk management team evaluates and adjusts trading strategies, providing assessment reports to the Portfolio Manager for final decision.
• The Portfolio Manager approves/rejects the transaction proposal. If approved, the order will be sent to the simulated exchange and executed.

Clone TradingAgents:

git clone https://github.com/TauricResearch/TradingAgents.git
cd TradingAgents

Create a virtual environment in any of your favorite environment managers:

conda create -n tradingagents python=3.12
conda activate tradingagents

Install the package and its dependencies:

Alternatively, run with Docker:

cp .env.example .env  # add your API keys
docker compose run --rm tradingagents

For local models with Ollama:

docker compose --profile ollama run --rm tradingagents-ollama

TradingAgents supports multiple LLM providers. Set the API key for your chosen provider:

export OPENAI_API_KEY=...          # OpenAI (GPT)
export GOOGLE_API_KEY=...          # Google (Gemini)
export ANTHROPIC_API_KEY=...       # Anthropic (Claude)
export XAI_API_KEY=...             # xAI (Grok)
export DEEPSEEK_API_KEY=...        # DeepSeek
export DASHSCOPE_API_KEY=...       # Qwen — International (dashscope-intl.aliyuncs.com)
export DASHSCOPE_CN_API_KEY=...    # Qwen — China (dashscope.aliyuncs.com)
export ZHIPU_API_KEY=...           # GLM via Z.AI (international)
export ZHIPU_CN_API_KEY=...        # GLM via BigModel (China, open.bigmodel.cn)
export MINIMAX_API_KEY=...         # MiniMax — Global (api.minimax.io)
export MINIMAX_CN_API_KEY=...      # MiniMax — China (api.minimaxi.com)
export OPENROUTER_API_KEY=...      # OpenRouter
export ALPHA_VANTAGE_API_KEY=...   # Alpha Vantage

For Azure OpenAI, copy .env.enterprise.example to .env.enterprise and fill in your credentials.

For AWS Bedrock, install the extra with pip install ".[bedrock]", set llm_provider: "bedrock", configure AWS credentials (environment variables, ~/.aws/credentials, or an IAM role) and AWS_DEFAULT_REGION, and use a Bedrock model ID, e.g. us.anthropic.claude-opus-4-8-v1:0.

For local models, configure Ollama with llm_provider: "ollama". The default endpoint is http://localhost:11434/v1; set OLLAMA_BASE_URL to point at a remote ollama-serve. Pull models with ollama pull <name>, and pick "Custom model ID" in the CLI for any model not listed by default.

For any other OpenAI-compatible server (vLLM, LM Studio, llama.cpp, or a custom relay), use llm_provider: "openai_compatible" and set the endpoint via backend_url (or TRADINGAGENTS_LLM_BACKEND_URL), e.g. http://localhost:8000/v1 for vLLM or http://localhost:1234/v1 for LM Studio. The model is whatever your server serves. No key is needed for local servers; set OPENAI_COMPATIBLE_API_KEY when the endpoint requires one.

Alternatively, copy .env.example to .env and fill in your keys:

Launch the interactive CLI:

tradingagents          # installed command
python -m cli.main     # alternative: run directly from source

You will see a screen where you can select your desired tickers, analysis date, LLM provider, research depth, and more.

TradingAgents works with any market Yahoo Finance covers, using the exchange-suffixed ticker. Company identity and the alpha benchmark resolve automatically per market.

• US: AAPL, SPY
• Hong Kong: 0700.HK · Tokyo: 7203.T · London: AZN.L
• India: RELIANCE.NS, .BO · Canada: .TO · Australia: .AX
• China A-shares: Shanghai .SS, Shenzhen .SZ (e.g. 600519.SS for Kweichow Moutai)
• Crypto: BTC-USD, ETH-USD

An interface will appear showing results as they load, letting you track the agent's progress as it runs.

We built TradingAgents with LangGraph to ensure flexibility and modularity. The framework supports multiple LLM providers: OpenAI, Google, Anthropic, xAI, DeepSeek, Qwen (Alibaba DashScope, international and China endpoints), GLM (Zhipu), MiniMax (global + China), OpenRouter, Ollama for local models, and Azure OpenAI for enterprise.

To use TradingAgents inside your code, you can import the tradingagents module and initialize a TradingAgentsGraph() object. The .propagate() function will return a decision. You can run main.py, here's also a quick example:

from tradingagents.graph.trading_graph import TradingAgentsGraph
from tradingagents.default_config import DEFAULT_CONFIG

ta = TradingAgentsGraph(debug=True, config=DEFAULT_CONFIG.copy())

# forward propagate
_, decision = ta.propagate("NVDA", "2026-01-15")
print(decision)

You can also adjust the default configuration to set your own choice of LLMs, debate rounds, etc.

from tradingagents.graph.trading_graph import TradingAgentsGraph
from tradingagents.default_config import DEFAULT_CONFIG

config = DEFAULT_CONFIG.copy()
config["llm_provider"] = "openai"        # e.g. openai, google, anthropic, deepseek, groq, ollama; openai_compatible covers any OpenAI-compatible endpoint (vLLM, LM Studio, llama.cpp, ...)
config["deep_think_llm"] = "gpt-5.5"     # Model for complex reasoning
config["quick_think_llm"] = "gpt-5.4-mini" # Model for quick tasks
config["max_debate_rounds"] = 2

ta = TradingAgentsGraph(debug=True, config=config)
_, decision = ta.propagate("NVDA", "2026-01-15")
print(decision)

See tradingagents/default_config.py for all configuration options.

TradingAgents persists two kinds of state across runs.

The decision log is always on. Each completed run appends its decision to ~/.tradingagents/memory/trading_memory.md. On the next run for the same ticker, TradingAgents fetches the realised return (raw and alpha vs SPY), generates a one-paragraph reflection, and injects the most recent same-ticker decisions plus recent cross-ticker lessons into the Portfolio Manager prompt, so each analysis carries forward what worked and what didn't.

Override the path with TRADINGAGENTS_MEMORY_LOG_PATH.

Checkpoint resume is opt-in via --checkpoint. When enabled, LangGraph saves state after each node so a crashed or interrupted run resumes from the last successful step instead of starting over. On a resume run you will see Resuming from step N for <TICKER> on <date> in the logs; on a new run you will see Starting fresh. Checkpoints are cleared automatically on successful completion.

Per-ticker SQLite databases live at ~/.tradingagents/cache/checkpoints/<TICKER>.db (override the base with TRADINGAGENTS_CACHE_DIR). Use --clear-checkpoints to reset all of them before a run.

tradingagents analyze --checkpoint           # enable for this run
tradingagents analyze --clear-checkpoints    # reset before running


config = DEFAULT_CONFIG.copy()
config["checkpoint_enabled"] = True
ta = TradingAgentsGraph(config=config)
_, decision = ta.propagate("NVDA", "2026-01-15")

TradingAgents is LLM-driven, so two runs of the same ticker and date can differ. This is expected for a research tool built on language models, not a defect. The variation comes from a few distinct sources, and it helps to separate them.

Language model sampling is non-deterministic. Even at a fixed temperature, providers do not guarantee byte-identical output across calls, and reasoning models (the default GPT-5.x family, and any thinking-mode model) vary the most because their internal reasoning is itself sampled.

Live data moves. News, StockTwits, and Reddit return different content as time passes, so a run today sees different inputs than a run last week even for the same historical trade date. Pin the analysis date to hold the price and indicator window fixed, but the social and news sources still reflect "now".

To reduce variation you can lower the sampling temperature. Set temperature in your config (or TRADINGAGENTS_TEMPERATURE in .env); lower values make models that honor it more repeatable. The current curated models are reasoning-first and largely ignore temperature, so for tighter reproducibility use a non-reasoning model, which you can set explicitly via the Custom model ID option.

config = DEFAULT_CONFIG.copy()
config["llm_provider"] = "openai"
config["temperature"] = 0.0
# Reasoning models ignore temperature. For tighter reproducibility, set a
# non-reasoning deep/quick model explicitly (e.g. via the Custom model ID option).

What does not vary anymore: the analyzed company identity is resolved deterministically from the ticker before any agent runs, and the market analyst grounds exact price and indicator claims in a verified data snapshot. Earlier reports of "different companies" or fabricated price levels across runs are addressed by these two mechanisms.

Backtest results are not guaranteed to match any published figure. Returns depend on the model, the temperature, the date range, data quality, and the sampling above. Treat the framework as a research scaffold for studying multi-agent analysis, not as a strategy with a fixed, replicable return.

Contributions are welcome: bug fixes, documentation, and feature ideas; past contributions are credited per release in CHANGELOG.md.

Please reference our work if you find TradingAgents provides you with some help :)

@misc{xiao2025tradingagentsmultiagentsllmfinancial,
      title={TradingAgents: Multi-Agents LLM Financial Trading Framework}, 
      author={Yijia Xiao and Edward Sun and Di Luo and Wei Wang},
      year={2025},
      eprint={2412.20138},
      archivePrefix={arXiv},
      primaryClass={q-fin.TR},
      url={https://arxiv.org/abs/2412.20138}, 
}