Frequently Asked Questions

About 5 minutes for basic setup, 15 minutes to be fully productive.

1. Download and install (1 minute)
2. Grant microphone permissions (30 seconds)
3. Set your preferred hotkey (30 seconds)
4. Try the built-in Email or Slack context (2 minutes)

You can create custom contexts later, but the defaults get you productive immediately.

No technical skills required for basic use.

Download, install, and start using the built-in contexts immediately. The app is designed for professionals who want productivity gains, not a technical project.

For advanced users: Offline mode requires some setup, but it's optional. Cloud and BYOK modes work without any technical configuration.

You'd be surprised-most people adapt quickly.

A few options:

• Use a good microphone close to your mouth-you can speak quietly
• Step away briefly for longer dictation
• Use in private spaces (home office, car, private office)
• Start with shorter inputs-even 10-second dictations save significant time

Many users report that colleagues barely notice. And for remote workers or those with private offices, it's a non-issue.

It depends on which mode you choose:

• Offline transcription:No. Audio is processed entirely on your device. Zero network traffic.
• BYOK Mode:Audio goes directly to your chosen provider (Deepgram, OpenAI, etc.) via their encrypted API. It never touches Contextli servers.
• Cloud Mode:Audio is processed by our secure backend and immediately deleted after transcription.

Yes, and here's what they'll want to know:

• Local mode: Zero data transmission. All processing on-device.
• BYOK mode: Data flows to your organization's own API accounts.
• Cloud mode: Data processed by Deepgram (enterprise-grade, SOC 2 compliant).

We can provide security documentation, data flow diagrams, and compliance information. Contact [email protected] for IT/security reviews.

Local mode enables HIPAA-compliant use.

With complete offline mode, no patient data is transmitted. This is the safest approach for healthcare.

We are not a covered entity and don't offer a BAA, but the local mode architecture means no PHI ever reaches our systems. Consult your compliance officer for your specific use case.

Your data stays on your device. We have nothing to delete.

Transcripts are stored locally on your computer. If you uninstall Contextli, you can delete the local data folder.

We don't store your transcripts, audio, or content. There's nothing on our servers to delete-because we never had it.

For Cloud/BYOK modes: Any modern computer works fine. Minimal CPU/RAM usage.

For offline transcription:

• Mac with Apple Silicon (M1/M2/M3/M4) - Excellent performance
• Windows/Linux with NVIDIA GPU - Very good performance
• CPU-only - Works, but slower on long recordings

For offline AI processing: 8GB+ RAM recommended. 16GB+ for larger models.

● macOS - Full support, optimized for Apple Silicon

● Windows 10/11 - Full support including Auto-Paste

● Linux - AppImage/Deb available (community-supported)

Yes-any microphone your computer recognizes works with Contextli.

AirPods, Bluetooth headsets, USB microphones, built-in mics-all supported. Select your preferred input in Contextli settings.

Tip: Quality microphones closer to your mouth produce better transcription accuracy.

The app itself: ~200MB

With offline transcription: +1-3GB (depending on model size)

With offline AI processing: +4-8GB (depending on model choice)

For cloud or BYOK mode, you only need the base app (~200MB).

Yes, with Local mode configured.

Complete offline mode works completely offline. No network connection required.

Set this up before your flight-the models need to be downloaded while online.

Lifetime License ($99 one-time):

• Unlimited local transcription (no per-minute costs)
• Use your own API keys (BYOK) - you pay providers directly
• Offline AI processing for complete offline operation
• Lifetime updates included

Cloud Pro ($19/month):

• 10 hours of cloud transcription per month included
• No API keys or setup required - just works
• Access to premium voice models

Credits are our unit for cloud processing.

When you use cloud transcription or cloud AI, it consumes credits:

• ~1 credit per second of audio transcription
• Credits vary by AI model and response length

Important: BYOK mode and Local mode use zero credits-they use your keys or local processing.

Yes-with limitations.

We offer a free tier for personal use with limited cloud credits. Enough to try the product and use it occasionally.

For daily professional use, the Lifetime or Monthly plan removes limits.

Yes-contact us for volume licensing.

We offer:

• Team licenses (5+ users)
• Enterprise deployment options
• Custom terms for larger organizations

Email [email protected] for team pricing.

Templates require you to find them and fill in blanks. Contexts transform your speech automatically.

• Templates: Search for template → Select → Fill in placeholders → Edit
• Contexts: Press hotkey → Speak naturally → Formatted output appears

Contexts capture the "template" in how they transform your speech. You don't think about formatting-you just speak, and the output matches the context.

Yes-contexts can be exported and imported.

Create a context that works well for your team's use case, export it, and share. Team members import and use the same formatting rules.

This is useful for standardizing:

• Email formats
• Status report structures
• Bug report formats
• Customer communication style

As specific as you need.

Simple contexts might just say "format as professional email." Complex contexts can include:

• Specific formatting rules
• Tone and voice guidelines
• Structural requirements
• Examples of desired output
• Context about your role or company

The more specific your context prompt, the more consistent your output.

For documentation about code-yes. For writing code-not really.

Contextli excels at:

• PR descriptions and commit messages
• Code review comments
• Technical documentation
• README files
• Explaining code in plain language

It's not designed for dictating actual code syntax. For that, tools like Talon or GitHub Copilot are better suited.

If you can type in it, Contextli works with it.

Contextli doesn't integrate with apps-it works at the system level. Press hotkey → speak → output appears at your cursor, wherever your cursor is.

This means it works with:

• Gmail, Outlook, any email client
• Slack, Teams, Discord
• Jira, Asana, Linear, Trello
• Notion, Google Docs, Word
• Any text field in any app

No integrations to configure. No app-specific setup.

Yes, with considerations.

Transcription: Offline transcription supports 99+ languages. Cloud transcription supports 36 languages.

AI transformation: Works best with languages well-represented in AI training data. English, Spanish, French, German, and major languages work well. Less common languages may have variable quality.

Custom contexts: You can create contexts that output in specific languages or translate.

Three main differences: workflow, formatting, and friction.

ChatGPT:

1. Open browser tab (or app)
2. Navigate to ChatGPT
3. Type or speak your prompt
4. Wait for response
5. Copy the output
6. Switch to destination app
7. Paste
8. Edit for context

Contextli:

1. Press hotkey (from any app)
2. Speak
3. Output appears at cursor

Contextli is built for people who use AI dozens of times daily. The workflow is optimized for speed and minimal friction-not for exploration or complex conversations.

Phone dictation is transcription. Contextli is transformation.

Built-in dictation gives you exactly what you say-"ums," false starts, and all. You then spend time editing.

Contextli transforms your speech for the destination context. Same input, completely different output quality and format.

Plus, phone dictation doesn't have contexts, hotkey access, or work across all your apps.

Similar category, different approach.

Dragon focuses on highly accurate transcription and voice control of your computer.

Contextli focuses on transforming speech for communication contexts-email, Slack, tickets, etc.

If you need to control your computer by voice (accessibility, RSI), Dragon or Talon may be better. If you need to write faster, Contextli is designed for that.

It depends on your needs, but Contextli is designed specifically for professional communication.

Most dictation tools focus on transcription-converting speech to text. Contextli focuses on transformation-converting speech to formatted, context-appropriate output.

For emails, Slack, Jira, and everyday professional communication, Contextli's contexts make it faster than alternatives that require post-transcription editing.

Yes-about 3x faster for most people.

Average typing speed: 40-50 words per minute
Average speaking speed: 125-150 words per minute

But raw speed isn't the full picture. With transcription tools, you might speak fast but then spend time editing. With Contextli's transformation approach, the output is often ready to send with minimal editing-compounding the time savings.

Modern AI is remarkably good at understanding intent, not just words.

When you speak casually, Contextli's AI understands you want a professional email or a casual Slack message-based on the context you set.

It handles:

• Filler words ("um," "uh," "like")
• Self-corrections ("no wait, I mean...")
• Casual phrasing that needs professional tone
• Structure and formatting appropriate to context

Try it yourself-the results often surprise people.