# Kilo Code Coding agent for editor and terminal with a desktop app in beta. Projects, models, agent roles and tool permissions can be configured explicitly. Canonical profile: https://agenten-kompass.mmeierhoff.chatgpt.site/en/agents/kilo-code/ Editor: [Michael Meierhoff](https://agenten-kompass.mmeierhoff.chatgpt.site/en/about/) Sources reviewed: 2026-10-09 Basis: provider documentation. Examples are editorial suggestions, not hands-on product tests. ## Overview Editor extensions, CLI and Desktop belong to the Kilo family. Desktop uses the CLI agent core and adds workspaces as well as Python and Conda tools. The desktop surface is documented as beta. - Provider: Kilo - Product family: Kilo Code - Profile type: Coding agent with extensions, CLI and desktop beta - Environment: VS Code, JetBrains, CLI and Kilo Desktop Beta - Connections: Files, shell, Git, browser, model providers and Conda in Desktop - Setup: Install a surface, open a project and configure the gateway, your own keys or a local model server - Pricing model: Platform USD 0 or Teams USD 15/user/month; Kilo Pass from USD 19/month, inference/cloud separate ## Operation & oversight Operational details for this profile have not yet been researched. ## Provider Kilo Code develops an open platform for coding agents, model access and development automations. In its announcement published on October 6, 2026, Anaconda identifies Kilo as an acquired company and part of its AI Dev Factory. - Operator: Kilo Code, Inc. / Anaconda - Organization: Company - Website: https://kilo.ai - Reviewed: 2026-10-09 - Repository: https://github.com/Kilo-Org/kilocode - GitHub-Stars: 27540 (2026-10-09) - License: MIT - Provider source: https://kilo.ai/about - Provider source: https://www.anaconda.com/press/agent-swarms-red-team-agents-ai-dev-factory - Provider source: https://api.github.com/repos/Kilo-Org/kilocode ## Prices and included usage Pricing checked: 2026-10-09 The platform and model inference are billed separately. Free/local models, BYOK, gateway usage and Kilo Pass are different access options. ### Free & Open Source USD 0 monthly platform fee Individual use; free/local models or separately paid inference. Price source: https://kilo.ai/pricing ### Teams USD 15 per user per month Team administration and collaboration; inference additional. Price source: https://kilo.ai/pricing ### Kilo Pass From USD 19 per month Optional inference credit subscription with bonus credits. Price source: https://kilo.ai/pricing ### Enterprise Custom pricing Governance, security review and SLA commitments. Price source: https://kilo.ai/pricing Credit purchases have a 5% processing fee according to the table. Cloud compute is billed separately. The 14-day trial includes Enterprise features. Pricing details: https://kilo.ai/pricing ## Key features ### Edit code with local tools Desktop works with files, shell, Git and browser using the CLI agent core. Source: https://kilo.ai/docs/desktop/overview ### Organize desktop workspaces Desktop combines project folders, agent sessions and notebook work in a dedicated surface. Source: https://kilo.ai/docs/desktop/overview ### Plan, ask and debug Built-in agents such as Code, Plan, Ask, Debug and Architect structure the task. Source: https://kilo.ai/docs/desktop/quickstart ### Configure custom agents Markdown files under .kilo/agents define custom instructions and tool permissions. Source: https://kilo.ai/docs/customize/custom-modes ### Differentiate tool permissions Permissions distinguish allow, ask and deny for individual tools and actions. Source: https://kilo.ai/docs/getting-started/settings/auto-approving-actions ### Review pull requests automatically AI PR Reviews examines pull requests for bugs, security, performance and project-specific instructions. Source: https://kilo.ai/docs/automate ### Parallel agents in Agent Manager Agent Manager coordinates parallel agent sessions in VS Code with individual Git worktrees or a shared workspace. Source: https://kilo.ai/docs/automate/agent-manager ### External tools through MCP MCP connects the agent to additional data sources and tools. Source: https://kilo.ai/docs/automate ## Requirements and limits ### Desktop is beta Limitation: The standalone desktop app is marked as beta; scope and behavior can change between versions. Source: https://kilo.ai/docs/desktop/overview ### Account for Conda and local model management Requirement: The desktop quickstart requires a Kilo account for Conda and local model server management. Your own provider keys can be configured separately. Source: https://kilo.ai/docs/desktop/quickstart ### Choose autonomy before starting Requirement: According to the quickstart, Desktop starts with high autonomy. Select local work or a worktree and manual or automatic approvals before the task. Source: https://kilo.ai/docs/desktop/quickstart ### Task subagents share the workspace Limitation: Subagents delegated through the task tool work in their parent agent’s workspace. A task invocation does not create its own Git worktree. Source: https://kilo.ai/docs/automate/agent-manager ## Example workflows ### Prepare a change with Plan - Your input: A repository, desired behavior and suitable project checks. - Possible result: A plan followed by a code change in the selected workspace. - What to check: Check the working mode, tool permissions, diff and test results. ### Reuse a project role - Your input: A reviewed agent file with instructions and permitted tools. - Possible result: Repeatable handling of a clearly bounded development task. - What to check: Check the loaded role, permissions and result in the repository. ## Context The standalone desktop app is marked as beta; scope and behavior can change between versions. The desktop quickstart requires a Kilo account for Conda and local model server management. Your own provider keys can be configured separately. Official product documentation reviewed. Workflows and assessments are editorial; hands-on product tests are outside the scope of this directory. ## Sources - https://kilo.ai/docs - https://kilo.ai/docs/desktop/overview - https://kilo.ai/docs/desktop/quickstart - https://kilo.ai/docs/getting-started/settings/auto-approving-actions - https://kilo.ai/docs/customize/custom-modes - https://kilo.ai/docs/automate - https://kilo.ai/docs/automate/agent-manager [Sources and method](https://agenten-kompass.mmeierhoff.chatgpt.site/en/sources/)