Solid transcription accuracy (we tested a number of competitors and they were the best) with great customer service. Dylan and team go above and beyond.
This isn't specific to Assembly AI, but transcription accuracy in general (especially with the audio quality of phone calls) can be hit or miss. It just comes with the territory. However, we have been really impressed with how frequently Assembly AI pushes accuracy updates to its models.
Assembly AI transcribes our customers phone calls. Without a solid and reliable ASR vendor, we don't have a product to sell.
I've tested many stt api's (Google/AWS/IBM) to transcribe audio/video files and AssemblyAI consistently wins on accuracy (lowest WER on our audio files).
They also worked with us to get pricing much lower than other api's, and cut our bill from AWS by 50%. Highly recommend for other devs looking for a really simple api/high accuracy.
Along with the rest of the speech recognition industry, there is still a ways to go to get to human level accuracy.
Accurate transcription of phone calls and recordings and video interviews.
Changing our transcript provider to AssemblyAI was probably one of the smartest decisions we made. Not only the accuracy is better but they support cutting edge features that can empower your product to do more. We did a benchmark across multiple providers and they were the clear winner. The team is great, super responsive and does a great job keeping us informed about new features and updates.
Nothing that I can recall at the moment.
Clean and super accurate transcription + audio intelligence features that make our product shine!
One of the best things about AssemblyAI is how much more affordable and accessible it is compared to many other options on the market. The pricing is straightforward and budget-friendly, which makes it an excellent choice for both small developers and larger teams. Despite the lower cost, the transcription accuracy and feature set remain top-notch. The API is easy to implement, and the documentation is clear and helpful. It’s reliable, fast, and packed with features like speaker diarization and topic detection that are usually reserved for much more expensive platforms.
Currently there are some features not available to the European users but I believe these are in development.
We use it to transcribe conversations between brokers and clients, which ensures that key details aren’t missed and can be easily reviewed or referenced later. This is incredibly valuable for our brokers, who can focus on the conversation without needing to take extensive notes, then use the transcriptions to follow up with tailored advice or next steps.
It accurately transcribes and diarizes every single transcript I throw at it
I think this is a fantastic solution, there's not many issues I have with it. Maybe if they improve the diarization algo, sometimes it merges speakers together
It has made the process of transcribing my meeting audio so simple and straightforward. I have all of those transcription being fed to an Agentic RAG tool, which has unlocked crazy new business value for me. Retrieving intelligence and insights across all of my meetings, when my own brain can't remember it, seriously been a game changer.
