Tabnine, a company that creates an AI-powered assistant for software developers that predicts when code will be generated for certain apps and fills in the blanks with relevant snippets, released a new version of its time-saving application programming interface on June 15.
The company, which operates on a freemium-plus-service provisioning business model, recently announced that its API has surpassed 1 million users in 2022.
Based in Tel Aviv CEO Dror Weiss told VentureBeat that Tabnine’s software has a large memory for discovering reusable code in 31 languages, including Python, JavaScript, Java, C#, TypeScript, and other languages available today in both Visual Studio Code and IntelliJ IDEs.
“Our product has a really straightforward use case,” Weiss explained. “Your developer uses an integrated development environment (IDE) to write code, and her editor includes a top line. We look at the code as it’s being written in real time and ask the AI engine contextual questions in real time; we then propose sections like the next line, snippets, or a block of code that the developer requires. The app makes a split-second decision about if this is what they wanted to do and accepts the code. This happens the majority of the time with little to no modification in the procedure.”
‘Almost all of the code has already been written’ – Tabnine CEO
Weiss claims that 90% of all code has already been written and does not need to be rewritten.
According to Weiss, developers are now seeing 30% to 40% of their code generated automatically utilising the Tabnine API without them having to stray. “No code is modified from their regular workflow of creating code in their favourite editor,” he explained. “It works for all popular concepts or code editors, and all major programming languages.”
Tabnine has overhauled both its AI models and its methodology to adding new models into its platform in the last year, according to Weiss. Tabnine may now support any suitable AI model for offering help to developers, as opposed to a monolithic approach that requires a single AI model, according to Weiss.
Companies that have used devops in recent years have witnessed productivity increases in their software development and release processes, from infrastructure-as-code to continuous integration. However, because an explosion of frameworks, tools, languages, and APIs has only made a developer’s job tougher, those productivity gains have generally halted at the developer’s keyboard, according to Weiss.
AI offers a lot of potential for increasing developer efficiency. Weiss believes AI has the ability to achieve for developer productivity what continuous integration has done for the release process.
“Every day, we get incredible feedback from developers and teams that Tabnine’s AI has become a part of their workflow, allowing them to focus on higher-value activities,” Weiss said in a media advisory. “Our next-generation AI technology, with longer and better code completions, is a huge step forward for developer productivity, further establishing Tabnine as the leader in AI-powered software development.”
The new platform contains a dozen new AI models, all of which were built from the ground up using code (vs. models pretrained on text and retrofitted to learn code). Tabnine can contextually complete full lines and snippets of code with extraordinary precision, according to Weiss, because each model is tuned for a certain language or area.
Tabnine, which was founded in 2017, recently revealed that it has raised $15.5 million in capital from Qualcomm Ventures, OurCrowd, and Samsung NEXT Ventures, as well as current investors Khosla Ventures, Headline Ventures, Hetz Ventures, and TPY Capital. Tabnine has now raised a total of $32 million with this new round of funding.
According to Weiss, Tabnine claims to be the only AI-powered software development tool of its sort on the market. According to Weiss, the new financing will be utilized to speed up the development of Tabnine’s platform in order to expand support for more AI models and programming languages later this year.
