TLDR: Bengaluru-based NimbleEdge has introduced DeliteAI, a pioneering open-source platform that enables developers to run agentic AI and large language models directly on mobile devices, eliminating the need for cloud-based infrastructure and enhancing user privacy.
Bengaluru-based deep tech startup NimbleEdge has officially launched DeliteAI, an open-source, on-device agentic AI platform designed to operate entirely on smartphones without reliance on cloud infrastructure. This development allows developers to create, deploy, and manage fully customized AI-native experiences directly on user devices, presenting a privacy-centric alternative to conventional cloud-based AI systems.
DeliteAI is a full-stack, production-ready infrastructure optimized for performance, privacy, and portability. According to NimbleEdge, the platform supports the deployment of transformer models, large language models (LLMs), and multimodal AI, thereby removing the necessity for cloud GPUs and ensuring that user data remains exclusively on the device. The platform comprises three main components: a production-ready SDK with an optimized inference stack and an on-device Python runtime for orchestrating agentic workflows; a dedicated Agent Marketplace for discovering and integrating pre-built AI agents; and the NimbleEdge Assistant, described as the world’s first fully on-device conversational AI assistant with integrated productivity tools.
Varun Khare, founder and CEO of NimbleEdge, stated, “We believe AI shouldn’t sit miles away from us in data centres. With this launch, any mobile application can scale AI to billions of users while improving data safety and user privacy.”
Unlike existing tools such as Ollama, LM Studio, and AutoGPT, which are primarily designed for desktop environments, DeliteAI is natively built for smartphones. The platform includes a mobile-optimized inference runtime for resource-constrained devices and a Python engine that is compatible with both Android and iOS. NimbleEdge defines agentic AI as modular, tool-using workflows that can complete multi-step tasks based on user context.
DeliteAI supports integration with open models like Llama, Gemma, and Qwen, and it abstracts away hardware complexities by supporting runtimes such as ONNX, LiteRT, and ExecuTorch. The Agent Marketplace offers reusable, on-device agents for a variety of functions, including text summarization, voice interaction (TTS/ASR), and recommendations.
Neeraj Poddar, co-founder and CTO of NimbleEdge, commented, “For the first time, developers can bring state-of-the-art AI models to consumer devices, orchestrate them with Python, and deploy truly private AI agents at scale.”
Swaroop Kolluri, Managing Director at Neotribe Ventures and lead investor in NimbleEdge, added, “By taking an open-source approach to on-device orchestration, NimbleEdge is making this shift more accessible to the broader developer community.”
Also Read:
- Gnani.ai Unveils Inya.ai: Empowering Indian Businesses with No-Code Agentic AI for Voice and Chat Across Over 10 Indic Languages
- EdgeRunner AI Launches Public Beta for U.S. Military, Offering Secure On-Device AI Agents
NimbleEdge has indicated that DeliteAI has already been battle-tested, having powered AI/ML experiences across over 30 million devices in production for gaming, e-commerce, and productivity applications.
