An AI assistant app that interacts with all major models, maintains history, generates and recognizes images, uploads PDFs/Word/Excel files, runs code, makes function calls to models, supports Markdown formatting, and more. Fully written by Claude 3 Sonnet.
- Gemini Pro 1.5, Claude 3 and GPT-3.5
- Audio Input in Gemini
- Video Input in Gemini
- Custom GPTs (best to use with LlaMa-3)
- Email respond (with attachments, like PDFs, Word, etc)
- Memory (for Gemini & Claude) - any information about user which you agreed to persist
- Scheduling any action (hourly or daily) to your email
- YouTube video summary
- Image Input (png/jpeg) in Gemini and Claude
- Image Generation by SDXL
- Together AI models databricks/dbrx-instruct mistralai/Mixtral-8x22B-Instruct-v0.1, LLama-3, etc
- File upload PDFs, Word, Excel into the conversation
- Multiple chats
- Markdown formatting
- Edit your messages with response regeneration
- Python Code Run
- PDF export of all chats
- Mobile friendly, PWA, Android app, etc
- Web Tools - weather, stocks, email send, Telegram, web search, etc - with Gemini, Claude, OpenAI models
- Please ask for more features in Discussions
- Instantly run Python code: Just ask AI to write your code and let Python Code Run take care of execution. You can even create pictures or graphs - simply instruct AI to save them to a file. Just say Create a python program which will draw house and clouds and save to disk
- Effortlessly fetch real-time data: Easily access important information such as weather updates, stock prices, and the latest news - just ask Whats weather in my country?
- Automate your email and Telegram correspondences: Send emails effortlessly to yourself or others. Just say Send information to me
- If your task require some compute, Python program will be created and executed in the background. Just say Calculate first 100 Fibonacci numbers
- Get the latest news from Google News in your preferred language and receive them directly in your inbox. Keep up with industry trends and advancements. Just say Send me your comment about latest news
- Create PowerPoint presentations. Just say Create 10 slides for presentation "Comparison of CZC.cz and Alza.cz" then write a Python program that will save it to disk then run code and Save To resulting .pptx link
- Schedule action and send results to you. Just say schedule hourly action: generate haiku about spring
- Write an email to Assistant allchatonline.ai@gmail.com with any inquiry (attachments allowed)
You have to get some of those APIs and set environment variables (or put to .env file in the server folder, you can start with /server/.env.example, rename it to .env):
- GOOGLE_KEY - Google cloud project name
- CLAUDE_KEY - Anthropic Key (for Haiku)
- TOGETHER_KEY - for Together models
- JWT_TOKEN - any random string
- OPENWEATHER_API_KEY - for weather tools
- YAHOO_FINANCE_API_KEY - for stocks tools
- TELEGRAM_KEY - for Telegram tools to work
- EMAIL - for email tools
- EMAIL_PASSWORD - for email tools
- Start local Mongo DB (or containerized:
docker run -p 27017:27017 -d mongo
) - In the server folder,
npm i
thennpm run api
- In the root folder,
npm i
thennpm run start