FinRobot Documentation

You can deploy the FastAPI-based web interface using the provided deploy.sh script. This will automatically install dependencies and start the server.

Access the interface at http://127.0.0.1:8001 by default.

chmod +x deploy.sh
./deploy.sh start

Before running FinRobot Pro, you must configure your API keys in a config.ini file.

1. Copy the example configuration file to the active config path:

   cp finrobot_equity/core/config/config.ini.example finrobot_equity/core/config/config.ini

2. Edit finrobot_equity/core/config/config.ini and provide your keys under the [API_KEYS] section:
   • fmp_api_key: Financial Modeling Prep API key.
   • openai_api_key: OpenAI API key.
   • adanos_api_key: (Optional) Enables Retail Sentiment Insights.

Follow these steps to set up a FinRobot development environment using Conda and Pip.

1. Create and activate a virtual environment:

   conda create --name finrobot python=3.10
   conda activate finrobot

2. Clone the repository:

   git clone https://github.com/AI4Finance-Foundation/FinRobot.git
   cd FinRobot

3. Install the package: You can install the latest release from PyPI:

   pip install -U finrobot

   Or install directly from the local repository in editable mode:

   pip install -e .

You can deploy FinRobot Pro with a single command using the provided deployment script, which handles dependency installation.

Using deploy.sh:

chmod +x deploy.sh
./deploy.sh start

Manual Deployment (if deploy.sh fails): If the script does not work, follow these steps to set up a virtual environment and run the app manually:

python3 -m venv venv
source venv/bin/activate
pip install -r requirements-equity.txt
python run_web_app.py

Once running, access the interface at http://127.0.0.1:8001.

Deployment Commands Reference:

After installation and configuration, you can explore the following notebooks located in the tutorials_beginner and tutorials_advanced directories to see the agents in action:

Beginner Tutorials:

• agent_fingpt_forecaster.ipynb
• agent_annual_report.ipynb

Advanced Tutorials:

• agent_trade_strategist.ipynb
• agent_fingpt_forecaster.ipynb
• agent_annual_report.ipynb
• lmm_agent_mplfinance.ipynb
• lmm_agent_opt_smacross.ipynb

FinRobot Desktop is a native macOS application for equity research.

Requirements:

• Apple Silicon Macs (M1, M2, M3, or later). Intel Mac builds are not supported in this release.

Installation Steps:

1. Download FinRobot_0.1.0_aarch64.dmg from the latest releases.
2. Drag FinRobot into your Applications folder.

Troubleshooting macOS Security Error: Because the app is not yet Apple-notarized, macOS may report it as "damaged" on first launch. To fix this, run the following command in your Terminal:

xattr -cr /Applications/FinRobot.app

Before running the analysis or report generation, you must configure your API keys. Copy the template to a new config.ini file and populate it with your credentials.

Required keys:

• fmp_api_key: For financial data and market metrics from Financial Modeling Prep.
• openai_api_key: For LLM-powered text generation.

Optional key:

• adanos_api_key: Enables Retail Sentiment Insights (Reddit, X.com, and Polymarket data).

cp finrobot_equity/core/config/config.ini.template finrobot_equity/core/config/config.ini

[API_KEYS]
fmp_api_key = YOUR_FMP_API_KEY
openai_api_key = YOUR_OPENAI_API_KEY
adanos_api_key = YOUR_ADANOS_API_KEY

FinRobot requires configuration for OpenAI and various financial data providers. You must set up two specific configuration files in the root directory.

1. OpenAI Configuration

1. Rename OAI_CONFIG_LIST_sample to OAI_CONFIG_LIST.
2. Remove the four lines of comments within the file.
3. Add your OpenAI API key.

2. Financial Data Provider Configuration

1. Rename config_api_keys_sample to config_api_keys.
2. Remove the comments within the file.
3. Add your specific API keys for the following services:
   • finnhub-api: YOUR_FINNHUB_API_KEY
   • financialmodelingprep (FMP): YOUR_FMP_API_KEY
   • sec-api: YOUR_SEC_API_KEY (required for financial report generation)

The following environment variables control the behavior and authentication of the web application:

| Variable | Default | Description |
|:---|:---|:---|
| `GITHUB_CLIENT_ID` | — | GitHub OAuth client ID |
| `GITHUB_CLIENT_SECRET` | — | GitHub OAuth client secret |
| `FINROBOT_ADMIN_EMAIL` | `admin@finrobot.com` | Initial admin email |
| `FINROBOT_ADMIN_PASSWORD` | (random) | Initial admin password |
| `WEB_HOST` | `127.0.0.1` | Server bind address |
| `WEB_PORT` | `8001` | Server port |

To use FinRobot, you must configure your OpenAI, SecApi, and FMPApi credentials.

1. OpenAI Configuration: Rename your OAI_CONFIG_LIST_sample file to OAI_CONFIG_LIST and update the API keys. Use autogen.config_list_from_json to load it.
2. Financial API Keys: Rename config_api_keys_sample to config_api_keys and replace the placeholder keys with your actual SecApi and FMPApi keys. Use register_keys_from_json to load them into the environment.

Ensure you define a work_dir for intermediate file storage (e.g., reports and instruction files).

To set up a multi-agent workflow:

1. Use register_keys_from_json to load your API credentials from a JSON file.
2. Initialize an AssistantAgent with your llm_config.
3. Initialize a UserProxyAgent to manage code execution and termination logic.

from finrobot.utils import register_keys_from_json
from autogen import AssistantAgent, UserProxyAgent

# Load API keys
register_keys_from_json("../config_api_keys")

# Initialize Analyst
analyst = AssistantAgent(
    name="Market_Analyst",
    llm_config=llm_config,
)

# Initialize User Proxy
user_proxy = UserProxyAgent(
    "user_proxy", 
    code_execution_config=False, 
    max_consecutive_auto_reply=1,  
    is_termination_msg=lambda x: x.get("content", "") and "TERMINATE" in x.get("content", ""),
    human_input_mode="NEVER"
)