AllChat

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.

Demo
Features
Connect model to world with Web Tools
Run Locally
- Environment variables
- Run

PROD (with $4.99 subscription and Premium features: Avatar Builder, Shared Chats, Achievements, Artifacts and more)

https://allchat.online

Features

Gemini Pro 1.5, Claude 3.5 and GPT-4o
Audio Input in Gemini
Video Input in Gemini
Custom GPTs
Email respond (with attachments, like PDFs, Word, etc)
Memory - 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

Connect model to world with Web Tools (switch on WebTools option)

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)

RUN LOCALLY

Environment variables

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

Run

Start local Mongo DB (or containerized: docker run -p 27017:27017 -d mongo)
In the server folder, npm i then npm run api
In the root folder, npm i then npm run start

graph LR
   subgraph AllChat["AllChat"]
       subgraph Application["Application Services"]
           Backend["Backend<br>extender777/allchat-backend"]-->MongoDB
           Frontend["Frontend<br>extender777/allchat-frontend"]-->Backend
           PythonShell["Python Shell<br>python:3.12"]-->Backend
       end

       subgraph Data["Data Services"]
           MongoDB["MongoDB<br>mongo:latest"]
       end

       subgraph Monitoring["Monitoring Services"]
           Prometheus["Prometheus<br>prom/prometheus"]
           NodeExporter["Node Exporter<br>prom/node-exporter"]
           Grafana["Grafana<br>grafana/grafana"]-->Prometheus
           Grafana-->Loki
           Loki["Loki<br>grafana/loki"]
       end
   end

   subgraph Volumes
       MongoData["mongo-data"]
       PrometheusVolume["prometheus"]
       GrafanaVolume["grafana"]
       LokiVolume["loki"]
   end

   MongoDB-->MongoData
   Prometheus-->PrometheusVolume
   Grafana-->GrafanaVolume
   Loki-->LokiVolume

   classDef serviceName fill:#ffc107,stroke:#ff9800,color:#212121;
   class Backend,Frontend,PythonShell,MongoDB,Prometheus,NodeExporter,Grafana,Loki serviceName;

graph TD
A[AllChat API] --> B[Express Server]
B --> C[Routes]
B --> D[Middleware]
B --> E[Database]
B --> F[External Services]

    C --> C1[User Management]
    C --> C2[Chat Interactions]
    C --> C3[Tool Calls]
    C --> C4[Subscriptions]

    D --> D1[CORS]
    D --> D2[Rate Limiting]
    D --> D3[Authentication]
    D --> D4[Logging]

    E --> E1[MongoDB]

    F --> F1[AI Models]
    F --> F2[Email Service]
    F --> F3[Telegram Bot]
    F --> F4[Payment Gateway]
    F --> F5[Web Scraping]
    F --> F6[File Processing]

    F1 --> F1A[Anthropic AI]
    F1 --> F1B[Google Vertex AI]
    F1 --> F1C[OpenAI]

    F6 --> F6A[PDF Parsing]
    F6 --> F6B[Word Document Processing]
    F6 --> F6C[Excel Spreadsheet Handling]

    B --> G[WebSocket Server]

    H[Utilities] --> H1[Sitemap Generation]
    H --> H2[Cron Jobs]
    H --> H3[Image Processing]

msveshnikov / allchat

readme