Voice AI Platform

What do you like best about the product?

I love the built-in PII and credit card redaction features of smallest.ai because they work flawlessly in real time, automatically masking sensitive identity details during live conversations. This ensures immediate protection of customer data and saves our team significant time since we don't have to manually remove or hide personal or payment details. It's particularly beneficial for handling financial conversations where privacy and data security are crucial. The initial setup of smallest.ai was very easy and straightforward, allowing our team to integrate it quickly without any major issues. The platform has helped us automate PII redaction and improve compliance monitoring, which was a major reason for switching from a manual transcription and call-monitoring process. The ability to integrate smallest.ai with our internal compliance and customer management systems using its API to export call data and connect transcripts with our secure document storage has been incredibly efficient for our compliance team. Overall, my experience with smallest.ai has been very positive, and I appreciate its smooth real-time features.

What do you dislike about the product?

The reporting dashboard could be more flexible,. Exporting detailed call data for specific regional branches sometimes requires using the API.

What problems is the product solving and how is that benefiting you?

I use smallest.ai for quick deposition transcriptions, solving our backlog by automatically drafting texts. Its AI redacts PII, instantly protecting sensitive data, especially during financial discussions. It saved us time by replacing manual transcription and monitoring, improving compliance and security efficiently.

What do you like best about the product?

I love how smallest.ai instantly creates highly reliable initial text drafts which remove our production backlog, saving us days of manual transcription effort. The automatic word-level timestamps are incredibly useful as they match up perfectly with our legal audio tracks. I can click on any word in the transcript and the media line jumps straight to that exact millisecond, which really speeds up our deposition editing workflows. Word-level timestamps turn a static text transcript into a highly interactive, searchable, and easily editable media database, eliminating hours of manual timeline scrubbing and allowing legal teams to focus on strategy rather than video editing. The platform integrates natively with several real-time AI and communication tools like LiveKit, Pipecat, Twilio, and Zapier, which is great for streamlining voice and automation workflows. Its ultra-low latency and native integrations with frameworks are incredibly powerful for building real-time, conversational voice agents.

What do you dislike about the product?

I find myself relying heavily on manual API configurations because the developer dashboard currently lacks a user-friendly, visual interface for managing custom pronunciation dictionaries. The documentation for advanced features like setting up data redaction or handling rate limits is relatively sparse, which often means I need hands-on assistance from the support team. I also feel that the platform is lacking in terms of providing comprehensive performance metrics, as there is no built-in dashboard for tracking granular, language-specific Word Error Rates. Additionally, for workflows needing a wide range of pre-made or community voice options, the smaller, more curated voice library can feel restrictive. The limited webhook options also make managing complex, high-volume logging more difficult.

What problems is the product solving and how is that benefiting you?

I use Smallest.ai to instantly create highly reliable text drafts, eliminating our production backlog. The word-level timestamps turn my transcripts into interactive, searchable, and editable databases, speeding up deposition editing and letting us focus on legal strategy.

What do you like best about the product?

​The standout feature for our educational architecture is the engine's incredible handling of specialized, context-dependent text. Most traditional TTS systems stumble hard over technical scientific terminology or legal citations, requiring heavy SSML tagging to fix pronunciation. Smallest.ai's Lightning model natively infers the correct phonetic structure based on surrounding paragraph context, which has drastically cut down our manual audio editing hours during textbook localization workflows.

What do you dislike about the product?

While the core voice synthesis is top-tier, the advanced pronounciation dictionary manager in the developer portal feels a bit bare-bones. It functions perfectly via API overrides, but a more robust visual interface within the dashboard for managing custom lexicon tables would make large-scale asset tracking easier for non-technical editors.

What problems is the product solving and how is that benefiting you?

​We needed to scale our auditory accessibility compliance (ADA/Section 508) across hundreds of hours of web-based training modules without blowing past our quarterly production budget. Smallest.ai allowed us to automate high-fidelity, highly natural audio generation straight from our headless CMS pipeline, saving us thousands in traditional voice-actor studio booking fees while speeding up our deployment cycles by roughly 40%.

What do you like best about the product?

The build in PII and PCI redaction capabilities have been a major advantage for our team. Sensitive information such as card numbers, account details, and personal identifiers is automatically masked before transcripts are processed, eliminating the need for a separate redaction tool. This has streamlined our workflow, reduced operational overhead, and strengthened compliance.

What do you dislike about the product?

The initial on-boarding and redaction configuration could be more intuitive. While the support team was very helpful, additional documentation, implementation guides, and real world configuration examples would make deployment faster and easier for new users.

What problems is the product solving and how is that benefiting you?

Smallest.ai enables us to generate RBI complaint call transcript with automated PII/PCI redaction, removing the need for an additional compliance solution. This has simplified our technology stack, reduced governance effort, lowered operational costs and increased confidence among our legal and compliance teams

What do you like best about the product?

Was honestly surprised by the impressive language coverage of Pulse STT. It consistently handles regional calls in Spanish, French, Portuguese, and Hindi with strong accuracy under a single API wrapper instead of managing separate localized endpoints.

What do you dislike about the product?

Would appreciate more language-specific Word Error Rate (WER) reporting inside the default metrics dashboards. Right now, it takes extra script work on our end to accurately analyze local language performance.

What problems is the product solving and how is that benefiting you?

Pulse has unified our regional operations under one roof and eliminated severe transcription latency. Our user-facing voice agents finally feel production-ready because the lag from voice-to-text has been flattened.

What do you like best about the product?

I like that the live conversational feel of our voice agent improved significantly after switching to the Pulse feature of smallest.ai. It solves issues with conversational overlap and response lags that disrupt the immersion of voice bots. I appreciate that it helped us move past our product testing phase faster, as the core SST layer no longer stalls on user interruptions. With Pulse, there's no awkward pause after a user finishes speaking, making the conversational flow feel natural. Even our investors and test groups noticed it as an improvement during product demos. The initial setup was straightforward and surprisingly quick, allowing us to integrate Pulse into our voice agent stack in less than a day. The compliance-grade call logging, P11 redaction, and dispute evidence transcription features were also valuable, with P11 redaction saving us from onboarding another compliance vendor.

What do you dislike about the product?

I would like to see a wider variety of webhook options for streaming event logs. Our team also ran into some minor connectivity delays during heavy back-to-back testing sessions.

What problems is the product solving and how is that benefiting you?

It solves conversational overlap and response lags, hastening product testing with seamless user interruptions handling. With Pulse, our voice agent's flow improved, eliminating pauses after speaking, noted as a significant upgrade during demos.

What do you like best about the product?

Our team's ability to pass bearer tokens, handle WebSocket upgrades, and proxy requests without having to develop any new bridging software was a major advantage.

What do you dislike about the product?

It keeps our in-app voice search from bouncing off mobile users. Customers don't abandon their baskets because of delayed text matching because the text returns nearly instantly.

What problems is the product solving and how is that benefiting you?

Pulse fixes language mapping errors instantly during live clinical intake and prevents telehealth audio recording drops. With the automation of the raw note-taking process, we've significantly reduced post-call administrative work for doctors.

What do you like best about the product?

I appreciate the built-in PII and PCI redaction in smallest.ai, as it has been a huge win for us by automatically removing sensitive information like card numbers and account details before the transcripts reach our systems, eliminating the need for a separate masking tool. The compliance-grade call logging, along with dispute evidence transcription, is excellent for our needs. It provides a detailed and secure call record that enhances transparency and meets compliance requirements, especially with RBI-compliant call transcripts. The automatic PII redaction ensures the protection of sensitive customer information and reduces privacy and security risks. Additionally, smallest.ai offers impressive SOTA STT accuracy and better multilingual coverage, which were key reasons for switching from Cartesia. The initial setup of smallest.ai was amazingly easy.

What do you dislike about the product?

The onboarding docs around redaction setup could be clearer. We needed a few rounds with support to fine-tune configurations, so more examples in the documentation would definitely help.

What problems is the product solving and how is that benefiting you?

I use smallest.ai for PII and PCI redaction, ensuring card numbers and personal info are secure. It provides RBI-compliant transcripts without extra vendors and reduces governance overhead, making our legal team comfortable. Automatic redaction protects sensitive info, meeting compliance needs and reducing privacy risks.

What do you like best about the product?

I was surprised by the impressive language coverage that smallest.ai Pulse STT provides. It consistently handles regional calls in Spanish, French, Portuguese, and Hindi with strong accuracy under a single API wrapper. This has reduced development complexity and improved transcription accuracy across different languages. I love that this feature allows us to deliver a more consistent experience for our global users. We switched to smallest.ai because it offers better accuracy for clinical terms and more reliable real-time transcription, which significantly reduced the amount of manual corrections our physicians had to make after consultations. I also appreciate how smallest.ai integrates with our electronic health record (EHR) system, internal documentation platform, and meeting management tools. This integration allows transcriptions and clinical notes to flow directly into patient records, reducing manual data entry and making the overall documentation process much more efficient.

What do you dislike about the product?

The documentation around advanced configurations and concurrency handling could be improved. We also had to spend some extra time understanding certain implementation details, so more detailed examples and troubleshooting guides would make onboarding much smoother.

What problems is the product solving and how is that benefiting you?

I use Smallest.ai for real-time medical dictation and transcription. It ensures RBI-compliant transcripts and satisfies our legal team. The language coverage through a single API reduces complexity and development time, while direct EHR integration minimizes manual entry, streamlining the documentation process.

What do you like best about the product?

I really like the added call transcription and language change adaptations in smallest.ai. It's great that we didn't have to build a correction layer before showing it to customers. The language changes from English to different languages are transcribed perfectly, and the initial setup was really easy.

What do you dislike about the product?

Documentation is very thin. It took us hit and trail to figure this out so you can improve there.

What problems is the product solving and how is that benefiting you?

I use smallest.ai for call transcription, meeting our enterprise customers' demands with ease. Its language adaptations save time by perfectly transcripting English to other languages without requiring a correction layer.