Development Status :: 5 - Production/Stable
Bard API
GoogleThe python package that returns response of Google Bard through value of cookie.
Please exercise caution and use this package responsibly.
I referred to this github repository(github.com/acheong08/Bard) where inference process of Bard was reverse engineered. Using __Secure-1PSID
, you can ask questions and get answers from Google Bard. This package is designed for application to the Python package ExceptNotifier and Co-Coder. Please note that the bardapi is not a free service, but rather a tool provided to assist developers with testing certain functionalities due to the delayed development and release of Google Bard's API. It has been designed with a lightweight structure that can easily adapt to the emergence of an official API. Therefore, I strongly discourage using it for any other purposes.
hf-transllm
If you want to test the open-llama model, which is released under the Apache License (allowing free commercial use)
in various languages
, you can try using the hf-transllm package. hf-transllm also supports multilingual inference for various LLMs stored in hugging face repository.
Amazing Bard Prompts Is All You Need!
- Helpful prompts for Google Bard
Install
If you will not provide the language parameter (use english
, korean
, japanese
only as input text):
pip install bardapi
If you wish to use the Bard API, including various features:
pip install git+https://github.com/dsdanielpark/Bard-API.git
Due to certain dependency packages that are not compatible with 64bit windows(OS), we are releasing a lightweight alpha release of bard that only returns responses for simple requests. This release is a continuation of the pypi 0.1.18
version, which was maintained with lightweight and simple functionality. See alpha-release github branch for more details.
pip install bardapi==0.1.23a
Authentication
Warning Do not expose the
__Secure-1PSID
- Visit https://bard.google.com/
- F12 for console
- Session: Application โ Cookies โ Copy the value of
__Secure-1PSID
cookie.
Note that while I referred to __Secure-1PSID
value as an API key for convenience, it is not an officially provided API key.
Cookie value subject to frequent changes. Verify the value again if an error occurs. Most errors occur when an invalid cookie value is entered.
If you need to set multiple Cookie values
- Bard Cookies - After confirming that multiple cookie values are required to receive responses reliably in certain countries, I will deploy it for testing purposes. Please debug and create a pull request
Usage
Simple Usage
from bardapi import Bard
token = 'xxxxxxx'
bard = Bard(token=token)
bard.get_answer("๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค")['content']
Or you can use this
from bardapi import Bard
import os
os.environ['_BARD_API_KEY']="xxxxxxx"
Bard().get_answer("๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค")['content']
To get reponse dictionary
import bardapi
import os
# set your __Secure-1PSID value to key
token = 'xxxxxxx'
# set your input text
input_text = "๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค"
# Send an API request and get a response.
response = bardapi.core.Bard(token).get_answer(input_text)
Addressing errors caused by delayed responses in environments like Google Colab and containers. If an error occurs despite following the proper procedure, utilize the timeout argument.
from bardapi import Bard
import os
os.environ['_BARD_API_KEY']="xxxxxxx"
bard = Bard(timeout=30) # Set timeout in seconds
bard.get_answer("๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค")['content']
Further
Behind a proxy
If you are working behind a proxy, use the following.
from bardapi import Bard
import os
# Change 'http://proxy.example.com:8080' to your http proxy
# timeout in seconds
proxies = {
'http': 'http://proxy.example.com:8080',
'https': 'https://proxy.example.com:8080'
}
bard = Bard(token='xxxxxxx', proxies=proxies, timeout=30)
bard.get_answer("๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค")['content']
Reusable session object
You can continue the conversation using a reusable session.
from bardapi import Bard
import os
import requests
os.environ['_BARD_API_KEY'] = 'xxxxxxx'
# token='xxxxxxx'
session = requests.Session()
session.headers = {
"Host": "bard.google.com",
"X-Same-Domain": "1",
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.114 Safari/537.36",
"Content-Type": "application/x-www-form-urlencoded;charset=UTF-8",
"Origin": "https://bard.google.com",
"Referer": "https://bard.google.com/",
}
session.cookies.set("__Secure-1PSID", os.getenv("_BARD_API_KEY"))
# session.cookies.set("__Secure-1PSID", token)
bard = Bard(token=token, session=session, timeout=30)
bard.get_answer("๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค")['content']
# Continued conversation without set new session
bard.get_answer("What is my last prompt??")['content']
More features
Starting from version 0.1.18
, the GitHub version of BardAPI will be synchronized with the PyPI version and released simultaneously.
pip install git+https://github.com/dsdanielpark/Bard-API.git
- Multi-language Bard API
- Get image links
- ChatBard
- Executing Python code received as a response from Bard
- Translation to Another Programming Language
- Using Bard Asynchronously
- Bard Cookies
- Fix Coversation ID (Fix Context)
- Translation to another programming language
hf-transllm
Example code ofIn case the Google package is no longer available due to policy restrictions, here's a simple example code for using open-source language models (LLMs) in both English and multiple languages.
Usage
For the decoder models provided by Hugging Face, you can easily use them by either following a simple approach or overriding the inference method. You can explore various open-source language models at this link. Through the ranking information from Open LLM Leader Board Report repository, you can find information about good models.
other than English
For LLM that use languages from transllm import LLMtranslator
open_llama3b_kor = LLMtranslator('openlm-research/open_llama_3b', target_lang='ko', translator='google') # Korean
trnaslated_answer = open_llama3b_kor.generate("๋์ ๋ด ๋๋
๋ฐฐ๋ค์ด ์ข์ํ๋ ๋ด์ง์ค์ ๋ํด์ ์๋ ค์ค")
print(trnaslated_answer)
English
For LLM that use Refer https://github.com/openlm-research/open_llama or using like this:
from transllm import LLMtranslator
open_llama3b = LLMtranslator('openlm-research/open_llama_3b)
answer = open_llama3b.generate("Tell me about the Korean girl group Newjeans.")
print(answer)
Scripts
In the scripts folder, I have released a script to help you compare OpenAI-ChatGPT, Microsoft-EdgeGPT and Google-Bard. I hope they will help more developers.
Contributors
I would like to express my sincere gratitude for the contributions made by all the contributors.
License
MIT
I hold no legal responsibility; for more information, please refer to the bottom of the readme file. I just want you to give me and them a star.
The MIT License (MIT)
Copyright (c) 2023 Minwoo Park
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
Shifting Service Policies: Bard and Google's Dynamics
Bard's service status and Google's API interfaces are in constant flux. The number of replies is currently limited, but certain users, such as those utilizing VPNs or proxy servers, have reported slightly higher message caps. Adaptability is crucial in navigating these dynamic service policies. Please note that the cookie values used in this package are not official API values.
Bugs and Issues
Sincerely grateful for any reports on new features or bugs. Your valuable feedback on the code is highly appreciated.
Contacts
- Core maintainer: Daniel Park, South Korea
- E-mail: [email protected]
Reference
[1] https://github.com/acheong08/Bard
Important Notice
The user assumes all legal responsibilities associated with using the BardAPI package. This Python package merely facilitates easy access to Google Bard for developers. Users are solely responsible for managing data and using the package appropriately. For further information, please consult the Google Bard Official Document.
Caution
This Python package is not an official Google package or API service. It is not affiliated with Google and uses Google account cookies, which means that excessive or commercial usage may result in restrictions on your Google account. The package was created to support developers in testing functionalities due to delays in the official Google package. However, it should not be misused or abused. Please be cautious and refer to the Readme for more information.
Copyright (c) 2023 MinWoo Park, South Korea