General FAQ
What is Rephlo?
Rephlo is the AI workspace that lives inside every app you already use. Instead of switching to a browser tab, you select text and press a hotkey to run AI commands in place — and you also get multi-turn chat, document-grounded Knowledge Spaces, and screen-aware Visual Intelligence, all from your desktop. It runs on Windows, macOS, and Linux, works with 8 BYOK providers or a hosted option, and can run fully offline with an on-device model or Ollama.
Does Rephlo have a built-in AI chat?
Yes. Beyond running commands on selected text, Rephlo includes full multi-turn chat — ask follow-up questions, brainstorm, or work through a problem without leaving your current app. Your chats can draw on your Knowledge Spaces for document-grounded answers, and you can switch between AI providers or run offline with Ollama. See Chat & Conversations.
What is Visual Intelligence?
Visual Intelligence lets Rephlo see and reason about your screen. Capture any region — a chart, screenshot, error message, diagram, or UI design — and the AI explains it, extracts text from images, debugs error messages, or answers questions about what's on screen. It's part of Rephlo's AI workspace, alongside chat, Knowledge Spaces, and in-app commands. See the Visual Intelligence tutorial.
Is Rephlo free?
Rephlo itself is software.
- The App: Depending on the license, the app itself may be free or paid.
- The AI: You pay for the AI separately (unless using Local Models).
- BYOK: You pay OpenAI/Anthropic/etc. directly per use.
- Local (Ollama): Free (uses your electricity).
Is my data safe?
Yes.
- Local Models (Ollama): Your data never leaves your computer.
- Cloud Models (BYOK): Your data is sent directly to the provider (e.g., OpenAI) over an encrypted connection. Rephlo does not see, store, or train on your data.
- History: Stored locally in an encrypted database. Never synced to the cloud.
Does it work on Mac/Linux?
Yes -- Rephlo is fully cross-platform.
| Platform | Status | Notes |
|---|---|---|
| Windows | Full support | Primary release target |
| macOS | Full support | Active development target with 19 platform services |
| Linux | Full support | Global hotkeys, overlay, screenshot capture, system tray, and text replacement all work; advanced IME input for some languages may be limited |
macOS Users: The first time you launch Rephlo, you'll need to grant Accessibility and Screen Recording permissions in System Settings > Privacy & Security. The app will guide you through this automatically.
Can I share my Spaces with others?
Yes! Rephlo has built-in Space export and import:
- Export: In the Spaces view, select one or more Spaces and click Export Selected (or Export All). This creates a ZIP package containing Space metadata plus all associated data files.
- Import: Click Import in the Spaces view and select a ZIP package. Rephlo validates the package and imports the Spaces with all their data.
This makes it easy to share knowledge bases with colleagues or transfer Spaces between machines.
How do I backup my Commands?
Rephlo provides built-in export/import for Commands:
- Export: In the Commands view, use Export Selected or Export All to save Commands to a JSON file. You can also access these options from File > Export Commands.
- Import: Use Import in the Commands view or File > Import Commands to load Commands from a JSON file.
This is the recommended way to backup your custom commands and share them with others.
Can I export my usage history?
Yes. In the History view, click the Export CSV button to save your transaction history as a CSV file. This is useful for:
- Auditing and compliance
- Analyzing your AI usage patterns
- Cost tracking across providers
How do I report inappropriate AI output or get help?
Go to Settings > About > Contact & Support and click Report inappropriate content — this works whether you're signed in or not. If you're signed in, you can also use the avatar menu at the bottom of the sidebar: click your profile card > Help > Report inappropriate content.
Either path opens rephlo.app/contact in your browser with the report category pre-selected. The same Contact & Support panel also lists Website, Contact, and Email if you need general help instead.
Feature-Specific FAQ
How many providers can I configure?
There's no limit. You can configure multiple instances of the same provider (e.g., two OpenAI accounts with different models) or mix providers freely.
What's the difference between Commands and Chat?
- Commands: One-click automations. Select text > run command > get result.
- Chat: Multi-turn conversations for brainstorming, iteration, and complex tasks.
Use Commands for repetitive tasks (grammar fixes, translations). Use Chat for exploratory work.
Can I use Rephlo completely offline?
Yes, in two ways:
- On-device models — download one from Settings ▸ LLM Providers ▸ On-Device. Requires Pro, Enterprise, or a perpetual license.
- Ollama — install Ollama, download a model (e.g.,
ollama pull llama3), and configure Ollama as your provider.
Either way, disconnect from the internet and Rephlo continues to work. Before you rely on a local model for everything, read On-Device vs Cloud Models — they're capable, but meaningfully less so than a large cloud model.
What file types can Spaces read?
- Documents: PDF, DOCX, TXT, Markdown (.md), CSV
- Code: Any text-based code file (.cs, .py, .js, .json, .yaml, etc.)
- Logs: .log files
- Scanned PDFs: read via on-device OCR — see below.
- Note: A picture embedded inside an otherwise text-based PDF isn't separately extracted.
Does Rephlo read scanned documents?
Yes. If a PDF has no real text layer — it came from a scanner, a photo, or an export — Rephlo recognizes the text on your own computer, so the document becomes searchable and usable as context. Nothing is uploaded to be recognized.
It's on by default at Settings ▸ Privacy ▸ Read scanned documents on this device, and the first use downloads a ~15 MB recognition model once. If your documents are in a different language than the app, set Scanned document language to match.
For the full picture — attachments vs Spaces, scanned documents, and images — see How Rephlo Reads Your Files.
How do I use variables in commands?
Rephlo supports these placeholders:
{{input_content}}: Your selected text{{space_data_all}}: All data from the active Space{{filename}}: Content from a specific file in your Space (e.g.,{{style_guide_pdf}})
Can Commands access the internet?
No. Commands only use the text you provide and your Space context. The AI model may have training knowledge, but Commands don't fetch live web data.
Is there a token/character limit?
Limits depend on your chosen provider and model. Rephlo manages these limits automatically through Space compaction and smart truncation. Check your provider's documentation for specific context window sizes.