Tabnine

Tabnine is an artificial-intelligence code assistant that suggests completions, generates code from natural-language prompts and explains existing code inside a developer's editor. Founded in 2013 in Tel Aviv as Codota, the company took its name from a pioneering open-source completion tool and grew into one of the longest-running players in AI-assisted software development, distinguished by a focus on privacy and on-premises deployment.^[1]

History

The company was established in 2013 by Dror Weiss and Eran Yahav as Codota, initially shipping AI-based code completion for Java. The underlying Tabnine tool was created in 2018 by Jacob Jackson, a University of Waterloo student, and used early large-language-model techniques to complete code across many languages and editors. Codota acquired the product in December 2019 and subsequently adopted the Tabnine name.^[1][2]

Funding followed a steady path: a US$12 million Series A in April 2020 led by Khosla Ventures, US$15.5 million in June 2022 co-led by Qualcomm Ventures, and a US$25 million round in November 2023. In June 2023 the company added a chat agent for generating, explaining and testing code in natural language.^[1][3]

On 30 July 2026, Tricentis — an Austin-based test automation and quality engineering company — acquired Tabnine, adding its Enterprise Context Engine to Tricentis's agentic quality engineering platform. The deal marked the consolidation of AI code generation with automated software testing.^[4]

Technology and Approach

Tabnine's product runs from inline completion to multi-file agent tasks, but its distinguishing claim has always been deployment control. Customers can run the assistant on their own infrastructure, in a virtual private cloud or fully air-gapped, so that source code never leaves their network. The Enterprise Context Engine extends this by retrieving a team's own repositories, conventions and standards to ground suggestions in how that organisation actually writes software, rather than in generic public code.^[4][5]

That positioning targets regulated and security-conscious buyers — financial services, government, defence and large multinational engineering teams — for whom sending proprietary code to a third-party cloud model is the blocking objection, whatever the quality of the suggestions.

Applications or Impact

In practice Tabnine is used for boilerplate and test generation, code explanation during onboarding, refactoring of legacy modules, and enforcing house style across large teams. Its trajectory also illustrates the economics of the market: a specialist completion tool sustained itself for a decade before being absorbed by a quality-engineering platform, as standalone code completion became a feature rather than a product.

>Key Takeaways

• Tabnine began in 2013 as Codota and took the name of a 2018 completion tool it acquired in 2019.
• Its edge is deployment control: on-premises, private cloud and air-gapped installations for regulated buyers.
• Tricentis acquired it on 30 July 2026, folding code generation into agentic quality engineering.
• For Malaysian teams, the practical value is keeping source code inside national jurisdiction while still gaining AI assistance.

See Also

• GitHub Copilot
• Cursor (Code Editor)
• Code Generation
• Windsurf
• Vibe Coding
• PDPA (Malaysia)