Voice-to-Task

Capture tasks hands-free. Just speak naturally and Taskitos creates tasks with the right details extracted automatically.

How Voice-to-Task Works

Voice-to-Task uses a two-step AI process:

OpenAI Whisper converts your speech to text with high accuracy, even with accents or background noise.
GPT-4o-mini analyzes the text to extract: task titles, due dates, priorities, and recognizes multiple tasks in one recording.

All in Your Timezone

GPT receives your current date, time, and timezone. When you say "tomorrow at 3pm," it knows exactly what that means for you.

Using Voice-to-Task

Click the microphone icon

Found next to the Quick Add bar on your dashboard. On mobile, look for the mic button.

Grant permission (first time)

Your browser will ask for microphone access. Click "Allow" to proceed.

Speak your task(s)

Talk naturally. For example: "Remind me to call John tomorrow at 2pm about the project."

Click stop when done

Click the stop button or wait for automatic detection of silence.

Review and confirm

Taskitos shows the transcription and extracted tasks. Edit if needed, then save.

Tips for Best Results

Speak Clearly

Whisper is good with accents, but clear pronunciation helps. Avoid mumbling or speaking too fast.

Include Context

Say "Call John tomorrow at 2pm" instead of just "Call John." More context = better task extraction.

Use Sequence Words for Multiple Tasks

Say "First, call John. Second, email Sarah. Third, buy groceries." This helps AI separate multiple tasks clearly.

Reduce Background Noise

While Whisper handles some noise, a quieter environment gives better transcription accuracy.

Recording Multiple Tasks

You can dictate multiple tasks in a single recording. Taskitos will create separate tasks for each one.

Example Recording:

"I need to call the dentist tomorrow morning. Also, I should buy groceries this weekend. And remind me to email the quarterly report by Friday at 5pm."

Taskitos Creates:

• "Call the dentist" - due tomorrow 9:00 AM
• "Buy groceries" - due Saturday 9:00 AM
• "Email the quarterly report" - due Friday 5:00 PM

Review Before Saving

Always review extracted tasks before confirming. AI is good but not perfect - you might want to adjust a time or add more detail.

Supported Audio Formats

Voice-to-Task works with these audio formats:

M4A - iOS and Android default (recommended)
MP3 - Universal format
WAV - High quality, larger files
WebM - Browser recording default
OGG - Open source format
AAC - Advanced Audio Coding

Daily Limits

To keep costs sustainable while the service is free, there's a daily limit:

10 voice recordings per day

Resets at midnight in your timezone. Each recording costs ~$0.007 in API fees.

Make Each Recording Count

Record multiple tasks in one go instead of making separate recordings for each task. This helps you stay within the daily limit.

Next: Natural Language Dates

Learn how Taskitos understands natural language when you type task dates like "tomorrow" or "next Friday."

Natural Language Dates