Overview

• Agentic work: MORATUS does more than generate answers. It breaks a goal into steps, calls tools when needed (web search, file generation, connectors, and more), evaluates the results, and keeps going until the work is done.
• Model independence: The platform is not tied to a single language model. Models from different providers (the GPT, Claude, Gemini, Qwen, DeepSeek, GLM, Grok, and MiniMax families) are unified under one agent, tool, and product experience. Instead of picking a model, you pick a service tier (Auto, Fast, Expert).
• Shared account and context: Every client works with the same account, credit balance, memory, preferences, and projects.
• Continuity: With memory, projects, and scheduled tasks, work isn't confined to one chat; decisions, files, and recurring jobs carry over to later chats.
• You stay in control: Steps that change something in a connected app run only after your approval; on Desktop, the trust gate, approval modes, and checkpoints set the agent's boundaries (see Agent security).
• Output in any format: Alongside text answers, MORATUS produces Word, PDF, Excel, PowerPoint, images, video, music, podcasts, and working web projects.

The journey of a request

1. 1You write your request: “Can you research the electric vehicle market in Türkiye and prepare a short Word summary?”
2. 2MORATUS breaks the goal into steps: it will scan current sources first, then put the findings into a document.
3. 3It calls tools in turn: web_search to find sources, read_webpage to read the pages, and create_document to produce the Word document. The server validates and runs each call and adds the result to the chat.
4. 4It evaluates the result of each step; if something is missing, it plans the next step accordingly.
5. 5When the work is done, it summarizes the result in the chat; the document opens in the side panel and is added to the Library.

Each tool call appears as its own line while the reply is generated, so you can follow step by step which source MORATUS read and which file it produced (see How do tools work?). Credits are deducted based on actual model usage (see How are credits deducted?).

What can you do with MORATUS?

Frequently asked questions

• Do I need to choose a model? No. You choose a service tier instead of a model (Auto, Fast, Expert); MORATUS routes the request to a suitable provider (see Service tiers).
• Which devices can I use MORATUS on? On Web, on Desktop for Windows, and on Mobile for iOS and Android; see the Product family section below for details.
• Can I try it without an account? Yes, as a guest on Web (see Guest mode).
• Is it paid? The Free plan is free and includes 300 credits every month. The Expert model and creative generation are on the Plus and Max plans (see Individual plans).
• How is my data protected? Each user can access only their own records, model provider keys are kept on the server only, and your content isn't used to train models on MORATUS's behalf (see Data isolation).

Coming soon: MORATUS Speech (a voice experience that understands, organizes, and turns conversations into action) and MORATUS Snap (an experience that captures your screen, a document, or a moment and adds it straight into your MORATUS context).

Which product should I use when?

All three products use the same account, credit balance, memory, preferences, and projects, so you can pick up on one device where you left off on another. For details, see MORATUS Desktop and MORATUS Mobile.

In the table, ✓ means the feature is on that platform and — means it isn't. Your account, credits, memory, preferences, and projects are shared across all three platforms.

Which one should I use?

• Web: To try MORATUS without an account and to work from anywhere in the browser.
• Desktop: To work in local folders, let the agent use a terminal and a browser, and approve and undo file changes (see MORATUS Desktop).
• Mobile: To continue chats from your phone and for voice chat (see MORATUS Mobile).

Web, Desktop, and Mobile connect to the same MORATUS API. The API handles identity and data, model providers, and generation engines in one place, which is why your account, credits, memory, and projects are the same on every client.

1. 1Clients: Web (React, in the browser), Desktop (an Electron shell and a local runtime), and Mobile (React Native / Expo; one codebase for iOS and Android).
2. 2MORATUS API: Model routing, tool execution, billing, memory, scheduled tasks, and connector operations run in this layer. Web and Mobile use the same API.
3. 3Identity and data layer: Sessions are opened with OAuth 2.0 and PKCE. Thanks to row-level security (RLS), each user can access only their own records.
4. 4Model providers: Requests are routed to the right provider based on the selected service tier. Provider keys are kept on the server only.
5. 5Generation engines: Separate engines run for Word, Excel, PowerPoint, Podcast, and Deep Research. Long-running jobs run on a durable job queue, and progress is streamed live to the chat.
6. 6How Desktop differs: File tools run on your computer, executed by the local runtime. Chats and generated files are written to disk; sign-in, model usage, memory, and credits come from the shared server.

Responses are delivered as a stream. Text, tool calls, and progress updates arrive as separate event types and are shown separately in the interface. Web, Mobile Because generation continues on the server, it completes even if you close the tab; if the connection drops, the stream resumes where it left off.

For platform-specific architecture, see MORATUS Desktop and MORATUS Mobile; for the security layers in detail, see Security and Privacy.