GPTStonks Chatbot API¶

Welcome to the GPTStonks Chatbot API documentation! This API allows you to interact with a powerful financial chatbot built on top of the openbb framework. Whether you're a developer looking to integrate financial chat capabilities into your application or a trader seeking automated financial insights, this API is designed to provide you with a seamless and customizable experience.

Table of Contents¶

• GPTStonks Chatbot API
• Table of Contents
• Introduction 🌟
• Features 🚀
• Supported LLM Providers
• Supported Embeddings Providers
• Getting Started 🛠️
  • Prerequisites
  • Full deployment [api + frontend + db] (recommended)
  • Installation 🛸
• For Production Environments 🏭
• Usage💡
• API Endpoints🌐
• Configuration with environment variables ⚙️
• Contributing 🤝
• License 📃
• Disclaimer

Introduction 🌟¶

GPTStonks is a financial chatbot powered by LLMs and enhanced with the openbb framework. It provides natural language conversation capabilities for financial topics, making it an ideal choice for a wide range of financial applications, including:

• Learning about the financial markets
• Improving trading strategies
• Financial news analysis: sentiment, trends, etc.
• Customer support for financial institutions

This API allows you to integrate the GPTStonks financial chatbot into your projects, enabling real-time financial chat interactions with users.

Features 🚀¶

• Real-time Financial Chat: Engage in natural language conversations about financial topics.
• Customizable Responses: Tailor the chatbot's responses to suit your specific use case.
• Easy Integration: Built on FastAPI, this API is designed for straightforward integration into your application or platform.
• Extensive Documentation: Detailed documentation and examples to help you get started quickly.

Supported LLM Providers¶

• Llama.cpp: optimized implementations of the most popular open source LLMs for inference over CPU and GPU. See their docs for more details on supported models, which include Mixtral, Llama 2 and Zephyr among others. Many quantized models (GGUF) can be found in Hugging Face under the user TheBloke.
• Amazon Bedrock: foundation models from a variety of providers, including Anthropic and Amazon.
• OpenAI: GPT family of foundation models. For now only instruct versions are supported, such as gpt-3.5-turbo-instruct. Chat versions will be added soon.
• Vertex AI: similar to Amazon Bedrock but provided by Google. This integration is in alpha version, not recommended for now.

Supported Embeddings Providers¶

• OpenAI Embeddings: includes models such as Ada 2.
• Hugging Face: including providers such as BAAI, see their general embedding (BGE) model list.

Getting Started 🛠️¶

Prerequisites¶

Before you begin, make sure you have Docker installed on your system.

Full deployment [api + frontend + db] (recommended)¶

1. Clone this repository to your local machine:

git clone https://github.com/gptstonks/api.git

1. Check the .env.template file for the required environment variables. Modify the values as desired and save the file.

2. Run the docker-compose file:

docker-compose up

1. If you didn't change any port configuration, navigate to http://localhost:3000 to access the frontend.

2. (optional) If you want to use your OpenBB PAT to access resources, you can set it from the API Keys section (sidebar) in the frontend.

Installation 🛸¶

1. Set up environment variables by creating a .env file in the project directory with the contents specified in .env.template.

2. [Highly Recommended] Option 1: use the latest Docker image in ghcr.io.

docker run -it -p 8000:8000 --env-file .env ghcr.io/gptstonks/api:main

• Option 2: build from source.

• Install PDM.

• Clone this repository to your local machine:

git clone https://github.com/GPTStonks/api.git

1. Navigate to the project directory:

cd projects/gptstonks_api

1. Install the required dependencies:

pdm install --no-editable --no-self

1. Create openssl.cnf to allow legacy TLS renegotiation, needed for OECD data in OpenBB.

echo 'openssl_conf = openssl_init\n\
\n\
[openssl_init]\n\
ssl_conf = ssl_sect\n\
\n\
[ssl_sect]\n\
system_default = system_default_sect\n\
\n\
[system_default_sect]\n\
Options = UnsafeLegacyRenegotiation'\
> openssl.cnf

1. Start the API:

uvicorn gptstonks.api.main:app --host 0.0.0.0 --port 8000

Now your GPTStonks Financial Chatbot API is up and running!

For Production Environments 🏭¶

For production environments, additional steps are necessary to ensure security and stability:

Ensure that uvicorn is configured with SSL certificates for secure HTTPS communication.

Build the Docker image from source:

docker build -t gptstonks-api:v0.1_pro -f Dockerfile.pro .

Now you can run the Docker image with the following command:

docker run -it -p 443:8000 \
-v /etc/letsencrypt/live/api.gptstonks.net/fullchain.pem:/api/cert.pem \
-v /etc/letsencrypt/live/api.gptstonks.net/privkey.pem:/api/key.pem \
--env-file .env gptstonks-api:v0.1_pro

Usage💡¶

To use the GPTStonks Financial Chatbot API, send HTTP requests to the provided endpoints. You can interact with the chatbot by sending messages and receiving responses in real-time.

API Endpoints🌐¶

Check http://localhost:8000/docs once the API is started to access the endpoints' documentation.

Configuration with environment variables ⚙️¶

Contributing 🤝¶

We welcome contributions from the community! If you have any suggestions, bug reports, or want to contribute to the project, feel free to open issues or propose changes.

License 📃¶

This project is licensed under the MIT License - see the LICENSE file for details.

Disclaimer¶

GPTStonks Chat serves as an interface for accessing financial data and general knowledge. It is not intended to provide financial or investment advice.