From podcasters to businesses who record telephone conversations between customers and staff there are a host of industries and disciplines that require audio or video transcription.
“AssemblyAI”, founded by Dylan Fox, uses generative AI to create speech and speaker recognition models that they offer to developers in the form of an API. This model can be used to develop a variety of products for the market, the function of which is only limited by developers’ imaginations.
The AI model is guaranteed to be production ready, scalable, secure and easy to use no matter the type of application or the language used to build it. AssemblyAI offers a JSON formatted feedback to API calls that’s an industry standard and is, therefore, simple to use.
A number of useful resources are available on AssemblyAI to acquaint developers with the AI model and the API. The company offers excellent documentation for its model, which includes :
- Product overview
- Detailed walkthroughs
- Video tutorials
- Industry updates
In addition to its rich training material, AssemblyAI also keeps developers informed about version updates, new features, improvements & bug fixes on a regular basis.
Currently, the platform has 3 different models available for developers to choose between:
- Core Transcription
Priced at $0.00025/second, this plan allows developers to convert audio, video or speech into text. The option allows for speech recognition, language detection, grammar & spelling customization in 12 languages. The model also supports 49 different types of audio & video files.
- Audio Intelligence
Priced at $0.000583/second this plan offers everything available in the Core plan as well as : summarisation, Content moderation, sentiment analysis & topic detection.
- LeMUR
Priced at $0.000118/second this plan offers everything in the Core plan plus custom summaries, AI coaching and allows files of up to 10 hours in length.
AssemblyAI has raised over $63 million from the likes of Y-Combinator to develop & perfect their popular product.
