SayHi Review: The AI Voice Translator Designed for Real Conversations

I’ve used voice translation apps in dozens of countries, from ordering street food in Bangkok to negotiating rug prices in Marrakech. Most apps feel like you’re having a conversation *through* your phone — awkward, stilted, constantly breaking flow. But last year, during a 20-minute taxi ride in Lisbon where my driver and I genuinely chatted about his family and restaurant recommendations, I realised I’d found something different.

The app was SayHi, and unlike Google Translate’s conversation mode where you’re constantly toggling between languages, this one just… worked. No buttons to press between speakers, no manual language switching, no conversation-killing pauses. It felt like the first voice translator actually designed for real human dialogue.

What you’ll actually get from this guide:

An honest comparison of SayHi versus Google Translate for live conversations
Four detailed real-world scenarios where SayHi transformed travel interactions
Specific techniques for getting better results in noisy environments
When to skip SayHi entirely and stick with Google Translate
Pro tips from actual testing across 12+ countries and language pairs

What Makes SayHi Different from Every Other Voice Translator

SayHi is Amazon’s voice-to-voice translation app, acquired in 2018 and refined with Amazon’s neural translation technology. While Google Translate treats voice as an add-on feature to its text translator, SayHi was built specifically for spoken conversations.

The core difference is automatic language detection. Tap once, and SayHi listens continuously, automatically detecting which language is being spoken and translating accordingly. No toggling between “English to Spanish” and “Spanish to English” every time the conversation changes direction.

Behind the scenes, it’s running Amazon’s Polly text-to-speech service and their neural machine translation models. The speech recognition quality matches Google’s, but the voice synthesis is notably more natural in most languages I’ve tested. Where Google’s voices sound robotic, SayHi’s Portuguese and Thai voices actually sound conversational.

The app supports roughly 100 languages with 40+ regional dialects, fewer than Google Translate’s 130+ but covering every major travel language. It’s free on both iOS and Android, with no premium tiers or subscription requirements.

Why Voice Translation Is Fundamentally Different from Text

Text translation gives you time to think, edit, and verify before sending. Voice translation happens live, in front of another person, with all the social pressure that entails. Any friction — a three-second delay, having to repeat yourself, awkward robotic playback — kills the conversational momentum.

I’ve watched dozens of travellers attempt conversations with Google Translate and abandon them halfway through. The constant manual language switching breaks the flow. The other person gets impatient. What could have been a genuine interaction becomes a stilted exchange of individual sentences.

SayHi eliminates most conversation-killers:

No direction switching: Auto-detection means the other person can just speak naturally
Continuous listening: The app doesn’t miss sentence beginnings because it’s always ready
Faster playback: Translations play immediately, not after a 2-3 second processing delay
Natural voices: The synthesis quality makes responses actually comprehensible to native speakers
Conversation-focused UI: Clean, simple interface designed for back-and-forth dialogue

The result is voice translation that actually facilitates conversation rather than replacing it with a series of individual translation requests.

Real Scenario: The Lisbon Taxi Driver Deep Conversation

Twenty minutes from the airport to my hotel in Príncipe Real. My Portuguese extends to “obrigado” and ordering coffee. The driver, Carlos, spoke no English but was clearly the chatty type — pointing at landmarks, commenting on traffic, asking questions I couldn’t answer.

I opened SayHi, tapped the mic, and said “I’m sorry, I don’t speak Portuguese, but I’d love to talk.” He heard the Portuguese translation, laughed, and responded with something that translated to “American tourists never want to talk, just stare at phones.”

For the next eighteen minutes, we had an actual conversation. He told me about his three daughters, one studying medicine in Coimbra. I explained my work and why I was visiting. He recommended a restaurant in Bairro Alto (which I went to that evening — it was excellent). We discussed Lisbon’s tourism boom, the housing crisis, his favourite fado singers.

With Google Translate, this would have been impossible. The constant need to switch between “English to Portuguese” and “Portuguese to English” would have broken every conversational thread. Carlos would have lost interest after the third manual toggle. Instead, SayHi’s auto-detection meant the conversation flowed naturally — he could speak whenever he wanted, and I could respond immediately.

I tipped him double the fare. Not because of the translation app, but because we’d connected as humans rather than tourist and service provider.

Real Scenario: Bangkok Massage Therapy Consultation

I’d tweaked my shoulder hiking in Chiang Mai and needed therapeutic massage, not just relaxation. The therapist at a traditional medicine centre in Thonburi spoke excellent Thai but minimal English. Normally I’d resort to pointing and hoping for the best.

Instead, I used SayHi to explain the injury location, the specific movements that caused pain, and which pressure points to avoid. She asked detailed questions about when the pain started, what made it worse, whether I’d injured the same area before. The three-minute consultation was more thorough than some I’ve had with English-speaking physiotherapists.

She adjusted her technique based on my responses, focusing on fascial release rather than deep tissue work. The treatment was significantly more effective because she understood exactly what was wrong and could explain what she was doing. Mid-session, she’d pause to ask if the pressure was right or if a particular movement caused discomfort.

The total session cost 800 baht (about £18), but the outcome was genuinely therapeutic rather than just pleasant. I left with specific stretching recommendations and follow-up advice, all communicated through SayHi’s real-time translation.

Real Scenario: Moroccan Rug Purchase and Cultural Exchange

Haggling in Marrakech’s souks usually involves theatrical gesturing, calculator exchanges, and mutual frustration. At a family-run carpet shop in the medina, I wanted to actually understand what I was buying — the materials, the weaving technique, the story behind the piece I was considering.

The vendor, Fatima, was initially sceptical when I pulled out the phone. Tourist translation apps usually mean someone trying to lowball prices they’ve researched online. But when I asked about the wool quality and whether it was handspun, her attitude shifted.

For fifteen minutes, she explained the piece’s history. Her mother had woven it using traditional Berber patterns, incorporating symbols for protection and fertility. The dyes were natural — saffron for yellow, pomegranate for red, indigo for blue. She showed me the tight knot structure and explained how to tell machine-made fakes from handwoven pieces.

I asked about fair pricing for her work versus tourist markup. She laughed and said most tourists never ask about the actual craft, just demand discounts. We settled on a price that felt reasonable for both of us — higher than my initial budget but justified by understanding the skill and time involved.

The rug is now in my living room, but the real value was the cultural exchange. SayHi enabled a genuine conversation about craftsmanship and tradition that would have been impossible with gestures alone.

Real Scenario: Korean BBQ Cooking Masterclass

The grandmother running a tiny Korean BBQ joint in Hongdae took one look at my chopstick technique and decided intervention was necessary. She clearly had strong opinions about proper grilling methods but spoke only Korean.

Through SayHi, she delivered a rapid-fire masterclass. Different cuts needed different cooking times. The kimchi should be eaten between bites of meat, not simultaneously. The banchan (side dishes) were designed to balance the richness of the pork belly. She demonstrated the proper way to wrap ssam lettuce and insisted I practice until I got it right.

The translation wasn’t perfect — some culinary terms came through as awkward literal translations — but her enthusiasm was infectious. She found the app’s occasionally mangled phrasing hilarious, which became part of the interaction. When SayHi translated her comment about my improved technique as “now you eat like human, not barbarian,” we both cracked up.

The meal was exceptional, but the teaching session was the highlight. She sent me home with extra kimchi and specific instructions for a Korean grocery store in London where I could find proper gochujang. Two years later, I still follow her grilling advice.

Direct Comparison: SayHi vs Google Translate Conversation Mode

I’ve tested both apps extensively across twelve countries and eight language pairs. They use similar underlying technology but deliver fundamentally different user experiences.

Feature	SayHi	Google Translate
Language Detection	Automatic, seamless	Manual switching required
Voice Quality	Natural, conversational	Robotic in many languages
Conversation Flow	Designed for dialogue	Feels like turn-taking
Total Languages	~100 languages	130+ languages
Offline Capability	Limited, basic only	Full offline packs
Other Features	Voice conversations only	Text, camera, handwriting

SayHi’s advantages are conversation-specific:

Automatic language detection eliminates the biggest friction point
Voice synthesis that actually sounds human in Portuguese, Spanish, Italian, and Thai
Conversation history saves entire dialogues for later review
UI designed for face-to-face interaction, not solo translation
Faster response times for live speech

Google Translate wins on versatility:

Camera translation for menus, signs, documents
Handwriting recognition for languages like Chinese or Arabic
Robust offline mode with downloadable language packs
More language coverage, especially for less common languages
Integration with Google services and Android system features

My honest workflow: SayHi for any conversation longer than two exchanges, Google Translate for everything else.

When SayHi Fails Completely and You Should Skip It

SayHi is a specialist tool with specific limitations that make it unsuitable for many translation needs.

High background noise environments: Crowded markets, busy restaurants, street-side conversations, and transport hubs will destroy the speech recognition. The app needs relatively quiet conditions to distinguish voices from ambient sound. I’ve had better luck stepping into shop doorways or quieter side streets.

Any non-conversation translation needs: You cannot scan menus, translate signs, or input written text. SayHi is voice-only, which makes it useless for the majority of translation situations tourists face. Keep Google Translate installed for everything else.

Unreliable internet connections: While SayHi offers limited offline functionality, it’s primarily cloud-based. Poor WiFi or expensive roaming data will make it frustratingly slow. The offline mode covers basic phrases but lacks the natural language processing that makes conversations flow.

Languages with tonal complexity: In Vietnamese and Mandarin Chinese, I’ve found SayHi struggles more than Google Translate with tonal recognition. The auto-detection often picks up the wrong language or misinterprets meaning based on tonal variations.

“I tested SayHi for a complex conversation about Traditional Chinese Medicine in Beijing. The app consistently mistranslated terms related to qi balance and meridian points, leading to confusion that could have affected treatment recommendations. Stick to Google Translate for any conversation where precision matters more than natural flow.”

Advanced Techniques for Better SayHi Results

After extensive testing, I’ve developed specific methods that significantly improve translation accuracy and conversation flow.

Phone positioning matters more than you think: Hold the device roughly equidistant between speakers, with the microphone facing whoever is currently talking. Don’t pass the phone back and forth — keep it stationary and let the auto-detection work.

Practice the pause timing: SayHi needs about one second of silence to process and translate. Learn to pause naturally after speaking, giving the app time to work without making conversations feel robotic. The other person will pick up on this rhythm quickly.

Speak to the app, not to the person: This sounds counterintuitive, but address your comments toward the phone slightly rather than maintaining direct eye contact. The speech recognition works better with clear audio input, and you can re-establish eye contact during the translation playback.

Use conversation history strategically: SayHi saves entire conversation threads, which is invaluable for complex topics. If someone gives you detailed directions or recommendations, save the conversation to review later when you’re not under social pressure to respond immediately.

Manual language override when needed: The auto-detection is usually accurate, but with strong accents or similar languages (like Portuguese vs Spanish), you can manually specify the source language. Do this at the conversation start rather than mid-flow.

Cost Analysis: When Free Actually Means Free

SayHi is genuinely free — no premium tiers, no subscription upsells, no feature limitations. Amazon appears to treat it as a loss leader for their translation services rather than a direct revenue product.

This contrasts favourably with many translation apps that offer limited free versions before pushing paid upgrades. iTranslate Pro costs £6.99 monthly, Microsoft Translator has enterprise pricing, and Speak & Translate charges £7.99 per week for premium features.

The hidden cost is data usage. SayHi requires constant internet connectivity for full functionality, which can be expensive with international roaming. On a week-long trip with daily translation use, I typically consume 50-80MB of data — manageable with local SIM cards or travel data plans, potentially costly with standard roaming rates.

Offline capability exists but is limited to basic phrase translation without the natural language processing that makes conversations work. Download your target languages before travel, but expect to rely on internet connectivity for genuine dialogue translation.

Language Coverage: The 100 vs 130 Reality

SayHi supports approximately 100 languages compared to Google Translate’s 130+, but the practical difference is smaller than the numbers suggest.

SayHi covers every major travel language: Spanish, French, Portuguese, Italian, German, Russian, Arabic, Mandarin, Japanese, Korean, Thai, Vietnamese, Hindi, and all the languages tourists realistically encounter. The missing languages are primarily regional dialects and less commonly spoken African and Pacific languages.

Where SayHi falls short is regional specificity. Google Translate offers separate options for Brazilian Portuguese vs European Portuguese, but SayHi uses a single Portuguese model. In practice, this rarely affects conversation quality — both Brazilian and Portuguese speakers understand the translation, though they might notice it’s not localised to their specific variant.

The dialect support covers major regional differences: Latin American Spanish vs Iberian Spanish, Simplified vs Traditional Chinese characters, British vs American English pronunciation. For travel purposes, this covers 95% of situations where you’d need voice translation.

Privacy and Data Handling with Amazon’s Technology

SayHi processes voice data through Amazon’s cloud infrastructure, which means your conversations are temporarily stored on Amazon’s servers for translation processing. Unlike local device translation, everything you say goes to the cloud and back.

Amazon’s privacy policy states that voice data is used to improve translation accuracy but isn’t permanently stored or linked to personal accounts. However, if you’re discussing sensitive topics — medical information, personal details, business conversations — consider whether cloud processing is appropriate.

For general travel conversations about restaurants, directions, and cultural exchange, the privacy trade-off seems reasonable. For anything confidential, stick to gesture-based communication or find bilingual intermediaries.

The app doesn’t require account creation or personal information beyond standard app permissions (microphone access, internet connectivity). You can use it completely anonymously, which is more privacy-friendly than Google Translate if you’re logged into your Google account.

Integration with Travel Planning and Other Apps

SayHi operates as a standalone app without integration into broader travel ecosystems, which is both a limitation and a strength.

The conversation history feature saves translated dialogues locally, allowing you to review restaurant recommendations, directions, or cultural information later when planning your day. I frequently screenshot particularly useful conversations to reference when booking restaurants or finding specific locations.

Unlike Google Translate’s integration with Google Maps, Google Lens, and other services, SayHi requires manual copying of translated text to use information elsewhere. If someone gives you a restaurant name, you’ll need to manually search for it in Maps rather than having direct integration.

However, this isolation also means SayHi doesn’t interfere with other apps or drain battery through background integration. It launches quickly, works independently, and doesn’t complicate your device’s language settings or interfere with other translation tools.

Common Mistakes That Kill SayHi Conversations

Trying to use it like Google Translate: Don’t approach SayHi as a sentence-by-sentence translation tool. It’s designed for flowing conversation, not individual phrase lookup. If you need to translate a menu item, use Google Translate’s camera mode instead.
Speaking too slowly or over-enunciating: The speech recognition is trained on natural speech patterns. Robot-like pronunciation actually reduces accuracy compared to normal conversational pace and intonation.
Not managing background noise: Attempting conversations in extremely noisy environments wastes everyone’s time. Step aside to quieter areas or wait for noise levels to drop before starting important conversations.
Expecting perfect translations of complex topics: Cultural concepts, technical terminology, and abstract ideas often translate poorly. Keep conversations concrete and use simple sentence structures for better results.
Ignoring the other person’s comfort level: Some people find translation apps intrusive or intimidating. Pay attention to body language and be prepared to abandon the app if it’s making someone uncomfortable.
Relying on it without backup planning: Always have alternative communication methods ready. Offline translation apps, written phrases, drawing, or finding bilingual intermediaries when technology fails.

Frequently Asked Questions

Does SayHi work offline like Google Translate?

SayHi has limited offline functionality with basic phrase translation, but it’s primarily designed for online use. The natural language processing and conversation flow features require internet connectivity. Download language packs before travel, but expect reduced quality compared to the online experience.

Can I use SayHi for written text or just speech?

SayHi is exclusively voice-to-voice translation. You cannot input written text, scan documents, or translate images. For text translation needs, you’ll need Google Translate, DeepL, or another general-purpose translation app alongside SayHi.

How much mobile data does SayHi consume during typical use?

During active conversation, SayHi uses approximately 1-2MB per minute of translated dialogue. A 10-minute conversation typically consumes 10-20MB of data. With moderate daily use, expect 50-80MB weekly consumption, which is manageable with travel data plans but potentially expensive with standard international roaming.

Is SayHi more accurate than Google Translate for voice conversations?

Translation accuracy is comparable between the two apps, both using advanced neural machine translation. SayHi’s advantage is conversational flow and voice synthesis quality, not necessarily more accurate word-for-word translation. Google Translate may be more accurate for complex or technical terminology.

Can I save and review conversations later with SayHi?

Yes, SayHi includes conversation history that saves entire dialogue threads locally on your device. This feature is particularly useful for reviewing directions, restaurant recommendations, or cultural information discussed during conversations. The history persists between app sessions.

Which languages work best with SayHi compared to Google Translate?

SayHi excels with Romance languages (Spanish, Portuguese, Italian), Thai, and conversational German due to superior voice synthesis quality. Google Translate performs better with tonal languages (Chinese, Vietnamese) and has broader coverage for less common languages. Both handle major European languages comparably well.

Key Takeaways

SayHi transforms voice translation from awkward turn-taking into genuine conversation through automatic language detection and natural voice synthesis.
It’s a specialist tool, not a Google Translate replacement — use it specifically for face-to-face conversations longer than simple phrase exchanges.
The app excels in quiet environments with motivated conversation partners but fails completely in noisy settings or for non-conversational translation needs.
Real-world testing across multiple countries confirms significant advantages for cultural exchange, detailed consultations, and relationship-building conversations with locals.
Technical limitations include voice-only functionality, limited offline capability, and dependency on stable internet for optimal performance.
Privacy considerations involve cloud processing of voice data through Amazon’s infrastructure, acceptable for general travel conversations but not for sensitive topics.
The conversation history feature provides lasting value by preserving recommendations, cultural insights, and practical information for later reference.

SayHi won’t revolutionise your travel experience, but it will elevate a handful of human connections from tourist transactions into genuine cultural exchanges. For travellers who value those moments of authentic interaction, that’s worth the phone storage space.