GPTPortal

A feature-rich portal to chat with GPT-4, Claude, Gemini, Mistral, & OpenAI Assistant APIs via a lightweight Node.js web app; supports customizable multimodalit

What is GPTPortal?

A feature-rich portal to chat with GPT-4, Claude, Gemini, Mistral, & OpenAI Assistant APIs via a lightweight Node.js web app; supports customizable multimodality for voice, images, & files.

Overview of GPTPortal

What it is

GPTPortal is a self-hosted, local-first web application that provides a single chat interface for multiple large language model (LLM) providers, including OpenAI, Anthropic, Google, xAI, DeepSeek, Groq, Moonshot, Mistral, the entire OpenRouter catalog, and any OpenAI-compatible endpoint (such as Ollama, LM Studio, vLLM). Users bring their own API keys and run the application on their own machine with a single node server.js command. The application is deliberately lightweight, with no build step, no bundler, no frontend framework, and no database. The backend is modular Node/Express; the frontend is hand-written HTML, CSS, and vanilla JavaScript. It is designed to be a tool that users own and understand, not a black box to deploy and forget.

Key features

  • Multi-provider support: Curated core catalog of models across OpenAI, Anthropic, Google, xAI, DeepSeek, Groq, Moonshot, and Mistral, plus the entire OpenRouter catalog loaded live and cached, and custom OpenAI-compatible endpoints auto-discovered into the model picker.
  • Token streaming: Answers appear as they are generated, with separate reasoning streams for models that support it.
  • Markdown rendering: Output is rendered with syntax-highlighted code blocks (highlight.js), math (KaTeX), and per-code-block copy buttons, sanitized before DOM insertion.
  • Multimodal capabilities: Vision (attach an image), file uploads (text-based files), voice (speech-to-text and text-to-speech), and image generation (prefix a message with Generate:).
  • Compare view: Send one prompt to several models at once and read answers side by side, each with its own cost.
  • Presets and controls: Save and reapply bundles of system prompt, temperature, max tokens, reasoning effort, and verbosity; live context, token, and cost tracking; export conversations to self-contained HTML files.

Use cases

  • Local-first AI chat: Run a single chat interface on your own machine to interact with multiple LLM providers without juggling different web UIs.
  • Cost-effective experimentation: Use your own API keys and pay only for what you use, avoiding multiple chat subscriptions.
  • Development and testing: Developers can read, modify, and restart the entire application in seconds, making it a base for building custom AI workflows.
  • Comparing models: Send the same prompt to several models simultaneously to compare responses and costs.
  • Multimodal interactions: Attach images, upload files, use voice input/output, and generate images within the same chat interface.

Getting started

bashgit clone https://github.com/Zaki-1052/GPTPortalcd GPTPortalnpm installcp .env.example .env# Edit .env to add your API keysnode server.jsThen open your browser to the address shown in the terminal (default: http://localhost:3000).

License and pricing

The license is MIT. The pricing type is open source. Users bring their own API keys and are billed directly by each provider for exactly what they use; there is no subscription or middleman.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/zaki-1052gptportal.svg)](https://yunodb.com/tools/zaki-1052gptportal?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/zaki-1052gptportal?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/zaki-1052gptportal.svg" alt="Featured on YunoDB" height="20"></a>