Tabnine launches ‘org-native’ AI agent platform

Tabnine has launched the Tabnine Agentic Platform for AI-assisted software development with coding agents. The platform enables enterprise software development teams to ship faster while maintaining control over code and context, the company said.

With Tabnine Agentic, introduced November 5, developers get autonomous coding partners that complete workflows, not just code suggestions or completions, all aligned with an organization’s standards and security policies, Tabnine said. Powered by the Tabnine Enterprise Context Engine, Tabnine’s “org-native” agents understand the users’ repositories, tools, and policies and use these artifacts to plan, execute, and validate multi-step development tasks, Tabnine said. Agent tasks include refactoring, debugging, and documentation. The engine incorporates coding standards, source and log files, and ticketing systems. Tabnine agents execute complete coding workflows, offering security and context, according to Tabnine.

Tabnine Agents can use external systems and tools to adapt to new codebases and polices without retraining or redeployment. The engine combines vector, graph, and agentic retrieval techniques to interpret relationships across codebases, tickets, and tools, enabling Tabnine’s org-native agents to reason through multi-step workflows, the company said. Enterprise-grade benefits cited include:

• Agents can automatically adapt to new codebases and policies, with no retraining or redeployment required.
• Agents can act and iterate autonomously through coding workflows.
• Centralized control ensures oversight of permissions, usage, and context.
• Contextual intelligence provides awareness of internal repositories, ticketing systems, and coding guidelines.
• SaaS, private, VPC, on-premises, and air-gapped deployments are all available and meet enterprise security standards.