Open Assistant API

✨ An out-of-the-box AI intelligent assistant API ✨

Introduction

Open Assistant API is an open-source, self-hosted AI intelligent assistant API, compatible with the official OpenAI interface. It can be used directly with the official OpenAI Client to build LLM applications.

It supports One API for integration with more commercial and private models.

It supports R2R RAG engine。

Usage

Below is an example of using the official OpenAI Python openai library:

import openai

client = openai.OpenAI(
    base_url="http://127.0.0.1:8086/api/v1",
    api_key="xxx"
)

assistant = client.beta.assistants.create(
    name="demo",
    instructions="You are a helpful assistant.",
    model="gpt-4-1106-preview"
)

Why Choose Open Assistant API

Feature	Open Assistant API	OpenAI Assistant API
Ecosystem Strategy	Open Source	Closed Source
RAG Engine	Support R2R	Supported
Internet Search	Supported	Not Supported
Custom Functions	Supported	Supported
Built-in Tool	Extendable	Not Extendable
Code Interpreter	Under Development	Supported
Multimodal	Supported	Supported
LLM Support	Supports More LLMs	Only GPT
Message Streaming Output	Supports	Supported
Local Deployment	Supported	Not Supported

LLM Support: Compared to the official OpenAI version, more models can be supported by integrating with One API.
Tool: Currently supports online search; can easily expand more tools.
RAG Engine: The currently supported file types are txt, html, markdown, pdf, docx, pptx, xlsx, png, mp3, mp4, etc. We provide a preliminary implementation.
Message Streaming Output: Support message streaming output for a smoother user experience.
Ecosystem Strategy: Open source, you can deploy the service locally and expand the existing features.

Quick Start

The easiest way to start the Open Assistant API is to run the docker-compose.yml file. Make sure Docker and Docker Compose are installed on your machine before running.

Configuration

Go to the project root directory, open docker-compose.yml, fill in the openai api_key and bing search key (optional).

# openai api_key (supports OneAPI api_key)
OPENAI_API_KEY=<openai_api_key>

# bing search key (optional)
BING_SUBSCRIPTION_KEY=<bing_subscription_key>

It is recommended to configure the R2R RAG engine to replace the default RAG implementation to provide better RAG capabilities. You can learn about and use R2R through the R2R Github repository.

# RAG config
# FILE_SERVICE_MODULE=app.services.file.impl.oss_file.OSSFileService
FILE_SERVICE_MODULE=app.services.file.impl.r2r_file.R2RFileService
R2R_BASE_URL=http://<r2r_api_address>
R2R_USERNAME=<r2r_username>
R2R_PASSWORD=<r2r_password>

Run

Run with Docker Compose:

docker compose up -d

Access API

Api Base URL: http://127.0.0.1:8086/api/v1

Interface documentation address: http://127.0.0.1:8086/docs

Complete Usage Example

In this example, an AI assistant is created and run using the official OpenAI client library. If you need to explore other usage methods, such as streaming output, tools (web_search, retrieval, function), etc., you can find the corresponding code under the examples directory. Before running, you need to run pip install openai to install the Python openai library.

# !pip install openai
export PYTHONPATH=$(pwd)
python examples/run_assistant.py

Permissions

Simple user isolation is provided based on tokens to meet SaaS deployment requirements. It can be enabled by configuring APP_AUTH_ENABLE.

The authentication method is Bearer token. You can include Authorization: Bearer *** in the header for authentication.
Token management is described in the token section of the API documentation. Relevant APIs need to be authenticated with an admin token, which is configured as APP_AUTH_ADMIN_TOKEN and defaults to "admin".
When creating a token, you need to provide the base URL and API key of the large model. The created assistant will use the corresponding configuration to access the large model.

Tools

According to the OpenAPI/Swagger specification, it allows the integration of various tools into the assistant, empowering and enhancing its capability to connect with the external world.

Facilitates connecting your application with other systems or services, enabling interaction with the external environment, such as code execution or accessing proprietary information sources.
During usage, you need to create tools first, and then you can integrate them with the assistant. Refer to the test cases for more details.Assistant With Action
If you need to use tools with authentication information, simply add the authentication information at runtime. The specific parameter format can be found in the API documentation. Refer to the test cases for more details. Run With Auth Action

Community and Support

Join the Slack channel to see new releases, discuss issues, and participate in community interactions.
Join the Discord channel to interact with other community members.
Join the WeChat group:

Special Thanks

We mainly referred to and relied on the following projects:

OpenOpenAI: Assistant API implemented in Node
One API: Multi-model management tool
R2R: RAG engine
OpenAI-Python: OpenAI Python Client
OpenAI API: OpenAI interface definition
LangChain: LLM application development library
OpenGPTs: LangChain GPTs
TaskingAI: TaskingAI Client SDK

Contributing

Please read our contribution document to learn how to contribute.

Open Source License

This repository follows the MIT open source license. For more information, please see the LICENSE file.

MLT-OSS/open-assistant-api