Bilingual AI workspace · EN / AR

One AI workspace for every workflow

Pervagans is a bilingual AI assistant and workspace. Chat with the best models — Claude, GPT, Gemini and more — then ground every project in your own instructions, files, skills and connectors. Run an AI agent, generate images and video, and get cited answers, all in one place.

Free plan with 500 monthly credits · No credit card to start · Web & desktop

Everything in one place

A complete AI workspace, not just a chatbot

From multi-model chat to autonomous agents and media generation — the capabilities that usually need five different tools, unified under one bilingual subscription.

💬

Multi-model chat

Talk to Claude Opus 4.8, Sonnet 4.6, GPT 5.5/5.4, Gemini Pro 3.1, GLM 5.1, DeepSeek V4 Pro and Qwen 3.7 Max — and pick the right model for each task.

🗂️

Spaces

Give each project its own pinned model, custom instructions, uploaded files and persistent memory so the assistant stays in context.

🎓

Skills

Toggle on expert Skills from a catalog of 58 — copywriting, deep research, data science, SEO, design — to inject real expert guidance into the model.

🔌

Connectors (MCP)

Connect Notion, Slack, GitHub, Linear, Google Drive, Gmail and more with one-click OAuth so the assistant works on your real data.

🎨

Image & video

Generate images with GPT Image, Qwen-Image and Wan, and video with Sora and Happy Horse — with aspect, seed and negative-prompt controls and a saved Gallery.

🔭

Deep Research

Autonomous multi-source research that returns a structured report with numbered citations you can trust and verify.

🤖

Agent Mode

An AI agent that plans and executes multi-step tasks by calling your connected tools, then writes a clear, final answer.

🌐

Bilingual EN / AR

A fully bilingual experience with complete right-to-left Arabic support — work and read cited answers in English or Arabic.

7+
leading AI models
58
expert skills
8
focused workspaces
2
languages (EN / AR)
How it works

Built to do real work, end to end

Each capability is designed to compound: chat picks the model, Spaces hold the context, Skills add expertise, and connectors let the agent act.

Chat with Claude, GPT, Gemini and more

Stop juggling subscriptions and browser tabs. Pervagans puts every frontier model behind one chat box, so you can choose the best brain for each task and compare answers instantly.

  • Claude Opus 4.8 & Sonnet 4.6, GPT 5.5/5.4, Gemini Pro 3.1, GLM 5.1, DeepSeek V4 Pro, Qwen 3.7 Max
  • Switch models mid-project without losing your context
  • Answers come back with numbered, verifiable citations
  • Canvas / Artifacts to edit long documents and code beside the chat
7+ frontier models · one chat · cited answers

Spaces and Skills that keep the AI in context

Organize work into Spaces — projects that remember. Pin a model, add instructions and files, and switch on the right expert Skills so every reply is grounded in how you actually work.

  • Per-project pinned model, custom instructions and persistent memory
  • Upload files once; the assistant keeps them in context
  • 58 reusable expert Skills you toggle on per workspace
  • 8 focused workspaces: Education, Writing, Translation, Data, Business, Financial, Image, Video
Pinned model + instructions + files + memory, per project

Connectors and an AI agent that gets things done

Connect your real tools with one-click OAuth over the Model Context Protocol, then let Agent Mode plan and execute multi-step tasks against your data — and report back clearly.

  • MCP connectors for Notion, Slack, GitHub, Linear, Google Drive, Gmail, Calendar and more
  • One-click OAuth — you authorize each connection
  • Agent Mode plans, calls your tools, and writes a final answer
  • Deep Research and Agent Mode combine for sourced, actionable output
Authorize → agent plans → calls your tools → final answer

Generate media and research, all in one workspace

Create on-brand images and video without leaving the chat, and run autonomous Deep Research that hands you a structured, cited report instead of a wall of links.

  • AI image generation with GPT Image, Qwen-Image and Wan
  • AI video generation with Sora and Happy Horse
  • Aspect, seed and negative-prompt controls, plus a saved Gallery
  • Deep Research returns a structured report with numbered citations
Image · video · Gallery · cited research reports
FAQ

Questions, answered

The essentials on models, pricing, connectors, Arabic support and privacy. See full plans on the pricing page.

What is Pervagans?
Pervagans is a bilingual (English and Arabic) AI assistant and workspace, available on web and desktop. You chat with leading AI models in one place, organize work into Spaces with pinned models and memory, enable expert Skills, connect your own tools through MCP connectors, run an AI agent that executes multi-step tasks, and generate images and video — all in a single privacy-first subscription.
Which AI models can I use in Pervagans?
You can chat with Claude Opus 4.8, Claude Sonnet 4.6, GPT 5.5 and 5.4, Gemini Pro 3.1, GLM 5.1, DeepSeek V4 Pro, and Qwen 3.7 Max — switching between them per task so you always use the right model for the job, with no separate subscriptions to manage.
Is Pervagans free to use?
Yes. The Free plan includes 500 monthly credits so you can try multi-model chat, Skills and connectors at no cost. When you need more, Pro is $20/month and Max is $60/month with 200,000 monthly credits. Annual billing saves roughly two months. Compare plans on the pricing page.
What are MCP connectors?
Connectors use the Model Context Protocol (MCP) to securely link external tools to the assistant with one-click OAuth — Notion, Slack, GitHub, Linear, Google Drive, Gmail, Calendar and more. Once connected, Agent Mode can read and act on your real data to complete tasks for you.
Does Pervagans support Arabic?
Yes. Pervagans is fully bilingual with a complete right-to-left (RTL) Arabic interface and English. You can work, chat and read cited answers in either language, making it a true bilingual AI assistant for Arabic and English speakers.
Is my data private?
Pervagans is privacy-first: your data stays yours. Files, instructions and memory live inside your own Spaces, connectors are authorized by you through OAuth, and answers come back with numbered citations so you can verify every claim.

Start building with one AI workspace

Chat with the best models, connect your tools, and put an AI agent to work — free to start, in English or Arabic.