Voice Input is a premium feature available to Starter, Pro, and Team subscribers. Each voice operation uses credits from your account.
Voice Entry Methods
Quick Voice Entry
Tap-to-speak with AI parsing and a visual confirmation dialog
Hands-Free Mode
Full conversational loop with audio waveform visualization - no screen touching required
Field Dictation
Dictate into individual text fields when typing isn’t convenient
Quick Voice Entry
Quick Voice Entry is the fastest way to log data by voice. Speak naturally, and AI extracts the relevant information from your speech.How It Works
1
Tap the voice button
Look for the “Add by voice” button on any helper that supports it. It appears near the add form.
2
Speak your entry
When the button turns red and shows “Stop”, speak naturally. For example:
“Spent forty-two dollars at Whole Foods on groceries”The button will pulse while recording.
3
Wait for processing
After you stop, you’ll see two processing states:
- “Listening…” - Transcribing your speech to text
- “Understanding…” - AI parsing your text into structured fields
4
Review and confirm
A confirmation dialog appears showing:
- Your raw transcription (exactly what you said)
- Parsed fields with extracted values
- Confidence badges on successfully filled fields
- “Held” warnings if required fields are missing
5
Add, edit, or cancel
Choose your action:
- Add - Confirms the entry as-is
- Edit - Opens the full form with pre-filled values
- Cancel - Discards the entry
Natural Speech Examples
Speak naturally - the AI understands context and extracts the right data:| What You Say | Fields Extracted |
|---|---|
| ”Coffee at Starbucks, five twenty” | Amount: 5.20, Description: Coffee, Merchant: Starbucks |
| ”Ran three miles in twenty-eight minutes” | Distance: 3, Duration: 28, Activity: Running |
| ”Call mom tomorrow at 2pm” | Task: Call mom, Date: tomorrow, Time: 2:00 PM |
| ”Feeling tired today, slept about 6 hours” | Mood: tired, Sleep: 6 |
Hands-Free Mode
Hands-Free Mode creates a full conversational loop between you and Peanuts - perfect when you can’t look at or touch your screen.
When to Use Hands-Free Mode
- 🚗 Driving - Log expenses or notes without touching your phone
- 🍳 Cooking - Track ingredients or recipe notes with messy hands
- 🏃 Exercising - Record workout data mid-session
- 👶 Multitasking - When your hands are literally full
The Conversation Flow
1
Enter Hands-Free Mode
Tap the headphones icon on supported helpers to launch the full-screen hands-free interface.
2
Tap to start
A large circular microphone button appears. Tap once to start the session.
3
Speak your entry
When you see “Listening…” and the audio waveform animation, speak your entry naturally.The waveform visualizes your voice in real-time using your helper’s theme color.
4
Listen to confirmation
Peanuts speaks back what it understood:
“Got it. Forty-two dollars at Whole Foods for groceries. Say yes to add, or no to cancel.”The screen shows your transcript and pending data.
5
Respond verbally
- Say “Yes” or “Confirm” to add the entry
- Say “No” or “Cancel” to discard
- Say “Edit” to exit and make manual changes
- Say “Exit” to close hands-free mode entirely
6
Continue or exit
After confirming, Peanuts says “Done” and waits for your next entry. The loop continues until you say “exit”.
Handling Missing Fields
If required fields are missing, Peanuts enters a “Held” state and prompts you:“I got forty-two dollars, but I need a category. What category is this?”You can then speak just the missing information, and Peanuts combines it with what it already has.
Voice Commands in Hands-Free Mode
| Command | Action |
|---|---|
| ”Yes” / “Confirm” / “Add it” | Saves the pending entry |
| ”No” / “Cancel” / “Never mind” | Discards the pending entry |
| ”Edit” / “Let me fix that” | Exits to manual editing |
| ”Exit” / “Close” / “I’m done” | Closes hands-free mode |
Peanuts uses a calm, utilitarian voice (ElevenLabs “Sarah”) designed for productivity - not a chatty assistant. Responses are brief: “Done”, “Got it”, “Held”.
Field Dictation
For simpler use cases, you can dictate into individual text fields without AI parsing.How Field Dictation Works
1
Find the microphone icon
Look for the 🎤 icon next to text input fields. Not all fields have this - typically longer text fields like notes or descriptions.
2
Tap and speak
Tap the microphone icon, speak your text, and tap again to stop.
3
Review transcription
The transcribed text appears directly in the field. Edit if needed.
Supported Field Types
Voice input works intelligently with different field types:| Field Type | How Voice Handles It |
|---|---|
| Text | Direct transcription or parsed extraction |
| Number | Converts spoken numbers (“forty-two” → 42) |
| Select | Matches spoken words to available options |
| Date | Understands “today”, “tomorrow”, “next Monday”, etc. |
| Checkbox | Interprets “yes/no”, “true/false”, “checked/unchecked” |
| Image | Skipped in voice entry - use form for photos |
Credit Usage
Voice features use credits from your account:| Operation | Credits Used |
|---|---|
| Speech-to-Text (transcription) | 10 credits |
| Voice Entry Parsing (AI extraction) | 1 credit |
| Text-to-Speech (hands-free responses) | 15 credits |
How are credits calculated for hands-free mode?
How are credits calculated for hands-free mode?
Each hands-free conversation step uses credits:
- Your speech → Transcription (10 credits)
- AI parsing → Understanding (1 credit)
- Peanuts response → Text-to-speech (15 credits)
What happens when I run out of credits?
What happens when I run out of credits?
If you run out of credits mid-session:
- Voice buttons show a lock icon
- An upgrade prompt appears
- You can still use manual text entry
Can I reduce credit usage?
Can I reduce credit usage?
To minimize credit usage:
- Use Quick Voice Entry instead of Hands-Free (no TTS credits)
- Batch multiple items in one spoken entry
- Use field dictation for simple transcription only
Browser Permissions
Voice input requires microphone access from your browser or device.Granting Microphone Access
- Chrome / Edge
- Safari
- iOS (Safari/PWA)
- Android (Chrome/PWA)
- Click the lock/info icon in the address bar
- Find Microphone in the permissions list
- Set to Allow
- Refresh the page
Troubleshooting
Voice button shows a lock icon
Voice button shows a lock icon
'No speech detected' error
'No speech detected' error
The transcription service didn’t detect any spoken words.Solutions:
- Speak clearly and at a normal volume
- Check that your microphone is working (test in another app)
- Move to a quieter environment
- Ensure the microphone isn’t covered or muted
Transcription is inaccurate
Transcription is inaccurate
Speech-to-text may struggle with:
- Heavy accents or unusual pronunciations
- Background noise
- Very fast or mumbled speech
- Technical jargon or proper nouns
- Speak slightly slower and more clearly
- Use the Edit option to correct mistakes
- For proper nouns, spell them out or edit after
AI parsing extracts wrong values
AI parsing extracts wrong values
If the AI puts data in the wrong fields:Solutions:
- Be more explicit: “The amount is forty-two dollars”
- Use field names from your helper in your speech
- Use Edit to correct, then confirm
- For repeated issues, consider renaming ambiguous fields
Hands-free mode doesn't hear my response
Hands-free mode doesn't hear my response
Solutions:
- Wait for the audio waveform animation before speaking
- Speak within 5 seconds (there’s a max recording duration)
- Tap the mic button to manually restart listening
- Check that TTS finished playing before responding
'Microphone access denied' error
'Microphone access denied' error
Your browser blocked microphone access.Solutions:
- Follow the Browser Permissions instructions above
- Check system-level microphone permissions (especially on Mac/iOS)
- Try a different browser
Best Practices
Speak Naturally
Don’t use robotic commands. Say “spent fifty bucks on dinner” not “amount fifty category food”.
Include Context
The more detail you provide, the better the AI parses. “Coffee at Starbucks” works better than just “coffee”.
Use Quiet Spaces
Background noise affects transcription accuracy. Find a quieter spot for better results.
Review Before Confirming
Always check the confirmation dialog. Voice entry is fast, but not always perfect.