Morana.

Chat with Morana, ask for explanations or work on a piece of writing together. Choose a configured AI provider to get started.

Add images using the picker, drag and drop or the clipboard. Whether Morana can interpret them depends on the selected model.

Search past conversations, rename them, pin important chats or archive them. You can also export and delete them.

Group related conversations into projects and start new chats there. Optional AI titles are generated after the first reply.

Use voice input and spoken replies. Depending on your setup, local speech options and online voices are available.

Configured voice activation can respond to “Morana” and start speech input. A microphone and wake-word model must be ready.

Morana’s visible presence includes facial expressions, eye movement and body motion. Her lips move in sync with spoken output.

Morana can remain visible outside the main window. You can show her only during audible replies and bring her forward when she speaks.

During voice input or output, Morana can lower other Windows audio sessions and restore their previous volume afterwards.

Connect selected files and folders as knowledge sources. The app stores their reference paths rather than copying the original files into app storage.

Knowledge processing supports formats including PDFs, Office documents, spreadsheets, email and code, as well as links and attachments. Results depend on the format, content and configured processing.

Personal context and learned preferences can persist across conversations. In settings, you can review entries, delete them or reset the active personality’s memory.

Ask Morana to forget or replace a learned preference. If the match is ambiguous, she asks which entry you mean.

Choose whether conversations contribute knowledge, how long it is retained and whether sources refresh automatically. This is controlled context, not unrestricted self-development.

In plan mode, Morana can prepare goals and steps, ask questions and request approval. Which steps can run depends on the available tools.

The plan panel shows progress, next steps, blockers and required input. From there you can return to the chat, review or stop.

Use enabled plugin and MCP tools in your conversation with Morana. Available actions depend on the connection, model and permitted scope.

Through a configured Codex connection, Morana can work on code and use system commands. Suitable tools, permissions and the active router are required.

Supported runs show commands, progress and limited live output in the conversation timeline. Technical details can be expanded when needed.

Morana provides connections for OpenAI, LM Studio, Ollama and other supported providers. Model capabilities and available tools differ.

OpenAI supports browser sign-in and API keys as separate authentication options. Usage and costs depend on the chosen access method.

Connect LM Studio or Ollama with suitable models. Using a local model does not make connected online services run offline.

Supported local models can use permitted MCP tools. The app takes the capabilities of the model actually loaded into account.

A chat can keep a separate session for each provider and resume it when you switch back. Supported models also offer adjustable reasoning levels.

Install supported signed plugins and enable or disable them. You can control whether Morana is allowed to access each plugin.

MCP connects additional services and tools. Morana provides server settings, connection status, sign-in and a list of available tools.

Search and select available plugin and MCP commands. For each MCP server, you can decide whether Morana may use it autonomously or only after your selection.

A suitable configured image service, for example through MCP, can handle image tasks. This extension has its own requirements and may incur separate fees.

The KeePass plugin can search a configured, unlocked .kdbx vault and look up entries. Access is read-only; it does not modify the vault.

Existing integrations include Gmail search and message viewing, plus Google Calendar sync and event creation. Drive uses a local DriveFS setup. The related dashboard interface is still in development.

Existing Windows integrations provide task and mail tools. Availability depends on Outlook or Microsoft Graph; message details and attachment downloads need the appropriate Graph connection. The dashboard is still in development.

The Moltbook plugin provides a feed, threads, inbox and direct messages. It requires a configured connection and the appropriate agent access.

Morana can use conversation and profile context to draft comments, her own posts and direct messages. By default, you approve sending.

Optional semi-autonomous and autonomous modes use an explicitly defined scope, consent, limits, pause and stop controls. This does not mean unrestricted background posting.

Chat-only mode disables AI tool actions. Standard mode uses approvals; privileged mode can run permitted actions without asking again. This does not replace missing credentials.

Supported plan and tool workflows show why a step is waiting, blocked or completed. Approvals and stops remain part of the workflow.

Self-checks inspect the app’s state and configuration. They distinguish detected issues from areas that could not be checked or are not configured.

Morana’s diagnostics chat can explain app settings and take you directly to them. Some safe changes show old and new values for confirmation; sensitive settings remain manual.

The app interface supports German, English and automatic system-language selection. Response language and speech settings are configured separately.

Control native Windows notifications with categories, quiet hours and optional grouping. Morana’s appearance, lighting and rendering quality can also be adjusted.

Desktop presence supports assisted, report-first and manager modes. What Morana may do remains bound by your operating mode, agreed scope and stop boundaries.