Anaconda, a provider of open-source tools for data science, machine learning and AI development that serves 45 million people globally, brought Python programming to business data analytics.
Now it has unveiled a desktop application called AI Navigator that lets users tap more than 200 pre-trained generative AI models and run them on your personal computer or local server – instead of sending data to the cloud or remote server – for tasks such as code generation and debugging.
We caught up with Peter Wang, co-founder and chief AI and innovation officer at Anaconda, to drill down on the AI Navigator.
The AI Innovator: What is the AI Navigator, and who is it for?
Peter Wang: AI Navigator is Anaconda’s desktop application that streamlines the use of AI models. It’s built for a wide range of users — whether beginners exploring AI or seasoned professionals seeking powerful tools. With access to over 200 curated large language models (LLMs), users can browse, download, and run models directly on their devices. What makes AI Navigator stand out is its simplicity and commitment to security — everything runs locally, ensuring your data stays private and on your machine.
What problem are you trying to solve with this solution?
Building and maintaining models and agents is highly technical and complex. Aligning the right model to the right problem is time-consuming for both creators and businesses. Testing solutions in the cloud can be costly, and running the latest Large Language Models (LLMs) on a personal device remains complicated and error-prone.
AI Navigator delivers open LLMs directly to the desktop, streamlining access and usability. Users can now run an integrated chatbot or inference server with over 200 curated models in just one click. Soon, we’ll introduce agent creation and management workflows, making it easy for users — regardless of skill level — to build agents tailored to their needs.
AI Navigator bridges a crucial gap between business and technical teams. While many AI tools cater to developers, we designed our solution to empower non-technical users to test models, offer feedback, and participate meaningfully in AI-driven initiatives.
Our roadmap focuses on simplifying AI agent creation and deployment, so teams can spend less time managing complexity and more time solving real-world problems. With AI Navigator and our upcoming AI Services, we aim to connect teams, making it easier to build, share, and integrate agents, models, and data seamlessly.
What tasks can developers do using AI Navigator?
AI Navigator empowers developers with easy access to Anaconda’s curated selection of the latest AI models. Using the integrated chat agent, developers can evaluate models and, by the end of the year, integrate them into applications with Anaconda’s services.
In 2025, developers can expect further expansion of tooling for AI agent creation, distribution, and deployment. A key advantage of AI Navigator is how it facilitates feedback from non-technical users on model performance. For example, an HR partner might use an AI-powered agent to check documents for compliance early in the process — helping developers fine-tune the model before wider deployment.
AI Navigator offers opportunities across multiple industries. Developers in finance can create agents for real-time market analysis and risk management, while healthcare professionals can design assistants for diagnostics and personalized treatments. In manufacturing, predictive maintenance systems and supply chain optimizers are possible. Customer service teams can build AI-driven virtual assistants, and government agencies can experiment with secure agents to improve public services.
The versatility of AI Navigator ensures that developers, regardless of industry or AI expertise, can unlock new possibilities and deploy advanced applications with ease.
How can AI Navigator help enterprises launch AI agents?
AI Navigator provides enterprises with secure, on-device access to LLM-powered chat assistants. For teams restricted from using cloud-based models due to compliance requirements, AI Navigator offers a local chat service that ensures proprietary data remains on employee devices, safeguarding sensitive information from third-party collection.
Looking ahead to 2025, Anaconda’s AI Services will enhance governance capabilities, enabling enterprises to control which models are used within their organization based on factors like model publisher and licensing. As AI Navigator integrates agent creation and distribution features, businesses will be able to deploy custom AI agents across departments to solve everyday challenges.
For example, customer service teams could build an on-premises AI agent fine-tuned with proprietary data, helping representatives quickly find information or even respond directly to customer inquiries. Software development teams could leverage AI Navigator to create coding assistants aligned with their unique codebase and practices, boosting developer productivity. Some enterprises are also using AI Navigator to develop productivity assistants that integrate with tools like Slack and email, summarizing conversations and managing information overload.
The key advantage of AI Navigator lies in its ability to securely implement AI solutions using internal data, without requiring extensive technical resources or large AI teams. This approach democratizes AI capabilities, empowering businesses of all sizes to innovate with ease.
Can you name the top 10 most downloaded models thus far?
- Meta-Llama-3-8B-Instruct
- Phi-3-Mini-4K-Instruct
- Codegemma-2b
- TinyLlama-1.1B-Chat-v1.0
- Meta-Llama-3-8B
- CodeLlama-13b
- Codegemma-7b
- Codegemma-7b-it
- CodeLlama-7b
- DeepSeek-Coder-7B-Instruct-v1.5
What are the most popular uses for your pre-trained LLMs?
Our usage data shows two primary trends. First, users are turning to our models as secure, local alternatives to cloud-based chat agents for tasks like ideation, summarization, and Q&A. These models offer the privacy and control that enterprises need, running entirely on user devices without exposing sensitive data to external servers.
The second major use case is in software development, where our LLMs are being used as coding assistants. Developers rely on these models to generate code, identify bugs, and streamline debugging processes, allowing them to complete tasks more efficiently and focus on higher-level problem-solving.
What results have you seen from your clients using AI Navigator?
AI Navigator has delivered notable results across industries by boosting productivity while ensuring regulatory compliance. In sectors like financial services and healthcare, where stringent data privacy rules restrict cloud-based AI tools, many organizations lack viable local alternatives. AI Navigator bridges this gap by enabling users to perform essential generative AI tasks on their own hardware, adhering to strict security and privacy policies.
The platform also accelerates individual users’ learning and adoption of generative AI. Thanks to its simplicity, AI Navigator has attracted a growing user base that includes students and professionals alike. Users appreciate the easy access to a diverse range of models and the intuitive interface, which makes interacting with AI tools more straightforward.
Early feedback from clients highlights the platform’s accessibility and ease of use. Users commend the seamless experience of deploying local models, along with the ability to experiment privately without compromising data security. Many have praised AI Navigator for its faster performance and simpler workflow compared to other solutions, making it a practical tool for organizations seeking to adopt AI without extensive technical resources. The design encourages collaboration between business and technical teams, helping enterprises integrate AI effectively and efficiently.
What are you doing to make sure AI Navigator is secure and private?
Once a user downloads a curated model to their device, all interactions with that model occur entirely on the local machine—no data from these interactions ever leaves the user’s computer. In fact, after the initial download, users can work with the models offline, with no need for an internet connection.
To further enhance security and efficiency, we perform model quantization through our secure infrastructure. This reduces model size, making them more suitable for various devices while mitigating risks associated with downloading from public repositories. Just as with Anaconda’s trusted open-source packages, we ensure that every model available through AI Navigator is genuine and uncompromised, safeguarding users from potential threats that can arise from tampered or malicious models.
What’s next for Anaconda?
We’re excited about the future at Anaconda. After a year of tremendous growth in 2024, we’re focusing on enhancing existing user workflows and unlocking new ones through agentic AI. By the end of the year, we’ll launch a unified intelligent platform powered by AI agents that automates and streamlines key processes, freeing users to concentrate on high-value tasks while we manage environments, dependencies, and package ecosystems in the background.
We are also integrating our products and technologies into a seamless, unified experience for users. Additionally, we’re developing AI agents to handle repetitive tasks and implementing telemetry tools to better understand how users engage with the platform. All these efforts align with our commitment to addressing real-world challenges in AI and data science, making these technologies more accessible and impactful for organizations.
Be First to Comment