# Help Your Team Code Faster and Explore Further with Kilo Desktop

> Source: <https://www.anaconda.com/blog/kilo-desktop-team>
> Published: 2026-11-17 14:39:18+00:00

Give developers and data scientists AI agents, notebooks, and model choice to get more done, with the controls managers need to scale adoption and outcomes.

Your developers want to ship useful software faster. Your data scientists want to explore more questions and turn data into reliable insights. Both are already using AI agents, but the work around the agent is split across editors, terminals, notebooks, and environment managers, and every handoff adds setup and review time.

For engineering and data science leaders, the opportunity is to increase what teams can accomplish while keeping review effort, AI spending, and administrative overhead manageable. That means reducing setup and repetitive work, making results easier to inspect, and giving teams model choice within clear organizational controls.

**Kilo Desktop by Anaconda is now generally available**, giving developers and data scientists an all-in-one application for their daily workflows. It brings together AI agents, access to more than 500 models, local inference, notebooks, and Python environment management. Kilo Teams and Kilo Enterprise add the visibility and administrative controls to support broader adoption.

Part of [Kilo by Anaconda](https://www.anaconda.com/products/kilo-by-anaconda), Kilo Desktop brings Anaconda’s [AI Workspaces](https://www.anaconda.com/platform/ai-workspaces) approach to the desktop, combining agentic tools, model choice, and trusted Python environments to help developers code faster and data scientists explore further.

With more than 5 million developers, processing 46 trillion tokens each month (as of October 2026), Kilo by Anaconda is a strong choice for teams that need model choice, open-weight models, data sovereignty, and cost control.

### Help developers move from planning to reviewed code faster

Writing code is one part of delivering software. Developers also need to understand unfamiliar projects, plan changes, manage dependencies, run tests, and review what their agents produce.

Kilo Desktop brings that workflow into one application. Plan mode reads the project and proposes an implementation approach before changes begin. Once approved, Code mode can divide the implementation into subtasks and run sub-agents in parallel. In Kilo, Workspaces group multiple folders so agents can reference the frontend, API, and shared packages together.

As developers delegate more work, session status indicators help them see which agents are working and which need input. Separate git worktrees give concurrent sessions their own working directories and branches.

Review happens in the same application as implementation. Developers can inspect diffs beside the conversation, run tests in a terminal, and preview an application in the built-in browser. When a change is ready, they can commit, push, and open a pull request from the app.

For engineering managers, the value is helping teams drive developer productivity by [moving 5-10x more work through coding, implementation, testing, and review](https://kilo.ai/customers/intraphone), with visibility into what AI agents changed and where developers need to intervene.

### Give data scientists more time for analysis and discovery

Data scientists need room to investigate. Following up a promising question can involve data preparation, exploratory analysis, several hypotheses, and repeated revisions before the findings are ready to inform a decision.

Kilo Desktop brings AI assistance directly into that process. Agents edit live Jupyter notebooks, so data scientists can review generated code alongside the cells, data, and outputs they are working with. They can ask for help exploring a dataset, investigating unexpected results, or extending an analysis while keeping the notebook in view.

Built-in conda environment management and access to Anaconda’s vetted packages from [AI Artifacts](https://www.anaconda.com/platform/ai-artifacts) help get the required Python dependencies in place. Code, notebooks, terminals, and agent conversations remain accessible in the same application as the investigation evolves.

For data science managers, less time on environment setup and repetitive coding leaves more time for testing assumptions, checking results, and answering business questions.

### Match models to the work and control spending

A complex coding task, an exploratory analysis, and a routine edit can have different requirements for model capability, speed, and cost.

Kilo Desktop gives teams access to [500+ models](https://kilo.ai/leaderboard), including OpenAI’s GPT models, Anthropic’s Claude, Google’s Gemini, and open-weight options such as NVIDIA Nemotron and Moonshot’s Kimi. Bring-your-own-key support and compatible local models give teams flexibility to choose the model, provider, and infrastructure that fit the work.

[Auto Efficient](https://blog.kilo.ai/p/auto-efficient) helps by routing requests to a lower-cost model that Kilo benchmarks show can handle the task. Support for local AI and local model inference gives teams another option for controlling where model processing happens. Our [local AI walkthrough](https://www.anaconda.com/blog/nvidia-rtx-spark-kilo-desktop-local-ai) shows how local models, Python environments, and notebooks work together.

Managers gain flexibility to support different workloads while making deliberate decisions about providers, infrastructure, and AI spending.

### Use the AI providers and subscriptions you already have

Teams building on Amazon Web Services (AWS) can use Kilo with models available through [Amazon Bedrock](https://kilo.ai/docs/ai-providers/bedrock), including Amazon Nova, in their coding and data science workflows.

Teams already using ChatGPT can connect an eligible subscription through [Sign in with ChatGPT](https://blog.kilo.ai/p/sign-in-with-chatgpt) and use its included model allowance in Kilo without a separate OpenAI API key. [ChatGPT Business support](https://blog.kilo.ai/p/announcing-business-plan-support) extends that option to teams, with administrators controlling access and how much of each member’s allowance Kilo can use.

Teams working in the Microsoft ecosystem can use Kilo with Microsoft Foundry, [VS Code](https://kilo.ai/landing/vs-code), and [GitHub](https://blog.kilo.ai/p/introducing-kilo-for-github). Anaconda’s trusted Python distribution and secure artifacts are also available to millions of Azure customers globally, supporting governed AI and data science development.

Teams can also bring their own keys (BYOK) from providers including Anthropic, Amazon Bedrock, Google AI Studio, OpenAI, Microsoft Foundry, Mistral, NVIDIA, and others. See the [BYOK guide](https://kilo.ai/docs/getting-started/byok) for supported providers and configuration options.

### Expand adoption with Teams and Enterprise controls

As usage grows, managers need to understand spending, share effective configurations, and manage access.

Kilo Teams adds centralized billing, shared agents and provider keys, and usage analytics. Leaders can see how their teams are adopting AI, while useful agent configurations can be shared across projects.

Kilo Enterprise adds model and provider restrictions, audit logs, single sign-on (SSO), and provisioning support. Organizations can establish approved choices and bring adoption into existing administrative processes. [Compare Kilo plans](https://www.anaconda.com/pricing) to find the appropriate level of support and control.

These capabilities expand Anaconda’s [AI Workspaces](https://www.anaconda.com/platform/ai-workspaces) approach, which brings together agentic tools and trusted building blocks while preserving model and tooling choice.

Organizations can extend protection with [Anaconda AI Security & Guardrails](https://www.anaconda.com/products/enkrypt-ai-by-anaconda) with [Enkrypt AI by Anaconda](https://www.anaconda.com/products/enkrypt-ai-by-anaconda) for closed-loop AI safety, combining AI supply chain security, security testing, runtime guardrails, and continuous compliance from development through production.

### Put Kilo Desktop to work on a real task

Start with a workflow your team knows well, such as implementing a feature or completing a recurring analysis. Compare time to a reviewed result, setup effort, rework, and model costs against your current approach.

[Download Kilo Desktop](https://kilo.ai/install) to start free, explore our [Kilo Desktop product walkthrough](https://www.anaconda.com/blog/kilo-desktop), or [contact Anaconda](https://www.anaconda.com/contact-us) to discuss your team’s requirements.
