Kilo Code
Also known as: Kilo, Kilocode, kilocode.ai
Open source coding agent across VS Code, JetBrains, CLI, cloud, mobile and Slack, with specialist agents delegating to parallel subagents across git worktrees, browser and terminal control, and 500+ models at zero markup. Acquired by Anaconda July 2026.
Kilo Code is an open source AI coding agent that runs across VS Code, JetBrains IDEs (IntelliJ, PyCharm, WebStorm), the command line, the cloud, and Slack. It takes instructions in plain language and carries out development tasks across several steps, including creating and editing files, running terminal commands, and testing in a browser. The project was forked from Cline and Roo Code and is now developed independently by Kilo-Org.
Its agents divide the work. Specialized agents cover Code, Ask, Plan and Debug, and any agent with full access can break a complex task into subtasks and delegate them to subagents, a capability built into every agent since the 2026 rebuild, when the older Orchestrator mode became legacy. Subagents can run in parallel across separate git worktrees with inline diff review, and you can approve each step or turn on auto approval for trusted workflows.
Model choice is central to the pitch. Kilo connects to more than 500 models through OpenRouter and direct providers, including Claude Opus 4.7, Claude Sonnet 4.6, GPT-5.5, and Gemini 3.1 Pro, all at provider rates with zero markup. You can switch models in the middle of a task, and an Auto Model feature routes each task to a suitable model.
The platform reaches well beyond code generation. Kilo Memory saves project facts, corrections and session digests and loads them in later sessions, and codebase indexing gives the agent semantic search across the whole project. Cloud Agents handle headless background execution, a dedicated Security Agent runs reachability analysis to prioritize exploitable vulnerabilities, and a Code Reviewer automates pull request review. Support for MCP servers adds custom tools.
Kilo Code was started by Sid Sijbrandij, a co-founder and former CEO of GitLab, and Scott Breitenother. It raised an $8 million seed round in December 2025 led by Cota Capital, with General Catalyst, Breakers, and Quiet Capital participating. The agent is used by millions of developers and ranks first on OpenRouter by coding token volume, and reported adopters include engineering teams at Meta, Amazon, Airbnb, PayPal, and Red Hat. The extension is free and open source. AI inference is billed separately, either by bringing your own provider key at cost or through Kilo Pass managed credits that start at $19 per month.
Vendor details
Canonical URL
https://kilo.ai
Category
Coding agent
Subcategory
Open source AI coding agent / agentic engineering platform
Funding status
Acquired by Anaconda Inc. on 15 July 2026 for an undisclosed sum, following Anaconda's earlier 2026 acquisition of Outerbounds. Kilo remains available with no announced changes to products, plans or support. Previously independent: cofounded by Sid Sijbrandij, GitLab cofounder and executive chair, and Scott Breitenother, who continues as CEO; raised an $8M seed round in December 2025 led by Cota Capital with General Catalyst, Breakers and Quiet Capital. Reported at over 3 million developers and ranked first on OpenRouter by coding token volume, with adopters including Meta, Amazon, Airbnb, PayPal, Square and Red Hat.
Company status
acquired
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Runs as a VS Code extension, a JetBrains plugin (IntelliJ, PyCharm, WebStorm), a CLI, in the cloud, and via Slack. Connects to 500+ models through OpenRouter and direct providers (Anthropic Claude, OpenAI, Google Gemini) using bring your own key at provider rates with zero markup, switchable in the middle of a task. Supports MCP servers for custom tools and GitHub for automated code review. Kilo Pass offers managed inference credits, and KiloClaw provides managed OpenClaw hosting as a separate product.
In practice
You're mid-feature and the task really needs planning, building, and debugging handled separately. Kilo's Orchestrator splits it into subtasks and routes each to a specialist mode instead of one agent fumbling through all three.
Your team is boxed into one model's pricing and limits. Kilo connects to more than 500 models at provider rates, and you can switch models in the middle of a task when one stalls.
The agent forgets your architecture every new session. Kilo's Memory Bank stores your decisions and conventions in repository markdown, so it rebuilds context instead of starting from scratch each time.
Sources & related URLs
Agentic Index coverage score
13.0 / 14 capabilities · 93%
| Integrations & Tool Calling | Full |
|---|---|
|
Runs as a VS Code extension, JetBrains plugin for IntelliJ, PyCharm and WebStorm, a CLI, a hosted cloud agent, iOS and Android mobile apps, and in Slack where the agent can be addressed directly to answer codebase questions or create pull requests; MCP servers install custom tools from a marketplace, and the agent controls the terminal and browser directly. Sourcekilo.ai/docs/code-with-ai and kilo.ai/coderead 2026-08-30 |
|
| Workflow Orchestration | Full |
|
Specialized agents cover Code, Ask, Plan and Debug with switching between them, and agents with full access delegate to subagents natively following the 2026 rebuild in which Orchestrator mode became legacy orchestration built into all agents; parallel tool calls and an Agent Manager shipped at GA in April 2026, subagents run concurrently across separate git worktrees with inline diff review, and task and todo lists break down complex work. Sourcekilo.ai/docs/code-with-ai and kilo.ai/coderead 2026-08-30 |
|
| Knowledge Grounding & RAG | Full |
|
Kilo maintains an index of the customer's code. Codebase Indexing, opt in per project or globally, parses code with Tree-sitter into semantic blocks, embeds them and stores the vectors in a vector database behind a semantic_search tool the agent calls across the whole project; AGENTS.md and context mentions add instructions and files. Sourcekilo.ai/docs/customize/context/codebase-indexingread 2026-09-29 |
|
| Human Oversight & Guardrails | Full |
|
A person approves each action by default: file changes and terminal commands are proposed with inline diffs and confirmation prompts before they apply, under one permission model shared by the VS Code extension, JetBrains plugin and CLI, with allow, ask and deny options per tool, and auto approve is an explicit opt in the docs warn against. Checkpoints restore earlier states. Sourcekilo.ai/docs/getting-started/settings/auto-approving-actionsread 2026-09-29 |
|
| Security, Identity & Governance | Full |
|
Enterprise documents SSO with OIDC and SCIM, owner and admin roles, audit logs of logins, invites, role and model changes, and limits on models and providers. Kilo's own trust subdomain, trust.kilo.ai, resolves to its parent Anaconda's trust center listing SOC 2 Type 2 reports, but no page states Kilo's systems are in that report's scope, so attestation for Kilo itself is not confirmed. Sourcekilo.ai/pricing, kilo.ai/docs audit logs and SSO pages, trust.kilo.airead 2026-09-29 |
|
| Observability & Auditability | Partial |
|
An AI Management Dashboard tracks adoption rates, usage, spend and velocity metrics across the organization with pooled credits and unified billing, in editor visibility shows what the agent is doing as it works, inline diff review exposes each change, and checkpoints save and restore working states. No retained per action execution trace or exportable audit log of agent actions has been found. Sourcekilo.ai/code and kilo.ai/docs/code-with-airead 2026-08-30 |
|
| Memory & State Persistence | Full |
|
Kilo Memory, opt in and scoped to one project, stores project facts, environment details, corrections and session digests that Kilo saves automatically from completed turns and loads in later sessions, kept until forgotten. The /memory forget command removes matching entries, /memory purge confirm deletes all of it, and an Inspect view shows the files. The older Memory Bank is deprecated and replaced by AGENTS.md, which holds instructions rather than memory. Sourcekilo.ai/docs memory pageread 2026-09-29 |
|
| Deployment & Data Residency | Full |
|
The extension is open source under a permissive license and runs locally in the developer's own editor or terminal, with bring your own key and support for self hosted and local models so code and inference need not leave the environment; a portable open source server built on OpenCode is shared across the extension, CLI and cloud agents, and self contained binaries including static builds for minimal containers are published. Sourcegithub.com/kilo-org/kilocode and kilo.ai/docsread 2026-08-30 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Skills ship as sharable packages of domain expertise, new capabilities and repeatable workflows with dedicated documentation, alongside specialized agents for Code, Ask, Plan and Debug that ship with the product, custom agents defined in.kilo/agents markdown files, and an MCP marketplace for installable server tooling. Sourcekilo.ai/code and kilo.ai/docsread 2026-08-30 |
|
| Triggers & Channel Coverage | Full |
|
Invocation spans VS Code, JetBrains, CLI, a hosted cloud agent, iOS and Android mobile apps, and Slack where the agent is addressed directly to create pull requests, with automated code review triggered on pull requests and headless autonomous runs in CI pipelines. Sourcekilo.ai/docs/code-with-ai and kilo.ai/coderead 2026-08-30 |
|
| Model Flexibility & Routing | Full |
|
More than 500 models from over 60 providers are reachable through the Kilo Gateway at provider rates with zero markup, or by bring your own key including self hosted deployments, switchable in the middle of a task; smart model routing automatically selects a model per task across efficient, frontier, balanced and free tiers with no manual switching, and named providers include Anthropic, OpenAI, Google, Mistral and Meta Llama. Sourcekilo.ai/code and kilo.ai/docsread 2026-08-30 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
Kilo can be extended in code: plugins receive a Kilo SDK client (@kilocode/sdk) for calling the local Kilo server along with lifecycle hooks and workspace adaptors, installable from npm or local folders, and the open source extension and CLI share that server; MCP servers, skills and custom agents extend it further. Sourcekilo.ai/docs plugins page and llms.txtread 2026-09-29 |
|
| Testing, Debugging & Optimization | Partial |
|
A Debug agent reads errors, traces issues and suggests fixes, Code Reviewer automates pull request review with line level comments, and a Security Agent runs reachability analysis to validate which vulnerabilities are exploitable against the codebase. These test the customer's code and review the agent's own output, and no harness to evaluate or regression test agent behavior is provided. Sourcekilo.ai/docs/code-with-ai and kilo.ai/coderead 2026-08-30 |
|
| Browser & Computer Use | Full |
|
Browser Use is listed as a core capability in the product documentation, letting the agent automate web interactions alongside terminal control, and the always on cloud agent can drive the browser headlessly; the extension also tests in a browser as part of its documented development loop. Sourcekilo.ai/docs/code-with-airead 2026-08-30 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
Several Kilo cloud agent chats can now work on the same branch at once, each with its own context, so one can plan while another implements. Each session also gets a diff viewer where line comments are gathered into a review and sent back to the chosen chat.
Bears on: Workflow orchestration
View sourceKilo Code CLI 7.7.9 added a goal tool that lets the agent start or resume a session goal and continue execution after the current turn. The tool is allowed by default, with ask and deny permission options. The same release added recurring or single session schedules through cron_create, cron_list, and cron_delete; tasks survive session resumption and expire after seven days.
Bears on: Human approval / guardrails
View sourceKilo's mobile app now lets users run cloud agents, control VS Code and CLI sessions, and review pull requests from iOS and Android.
Bears on: Workflow orchestration
View sourcePricing
Free and open source · Teams $15/user/mo · Enterprise custom · inference at provider rates, BYOK, or Kilo Pass from $19/mo
Free and open source to install and use, with model usage billed by consumption. You bring your own key and pay the model provider directly at zero markup, or route through the Kilo Gateway and pay published provider rates, and the optional KiloClaw always on cloud agent is a flat monthly subscription with inference billed separately.
Included quota
Free: the open source extensions and CLI with any model. Teams: usage analytics, shared agent modes, centralized billing, shared BYOK, team management. Enterprise: SSO, OIDC and SCIM, audit logs, model and provider limits, a shared private gateway and SLAs.
What is public
kilo.ai/pricing publishes the platform plans (Free for individuals, Teams $15 per user a month, Enterprise custom), inference choices (free models, local models, BYOK, Kilo Gateway pay as you go at exact provider rates, or Kilo Pass Starter $19, Pro $49 and Expert $199 a month with up to 50% bonus credits), a 5% credit purchase processing fee, and pay as you go cloud compute.
Billing mechanics
The platform plan is free for individuals or per user on Teams; inference is paid separately at provider rates through Kilo Gateway credits, through the customer's own keys, or through a Kilo Pass monthly credit subscription with bonus credits; cloud features add usage based compute.
Cost watchouts
Inference is billed by usage and grows with delegation to subagents; a 5% processing fee applies to credit purchases; cloud compute is separate.
Variable cost rationale
The software is free, so essentially all cost is model usage billed by consumption, and it scales directly with how much the agents run and which models they call. Orchestration and subagents multiply token use, and the cloud agent bills inference separately. Bring your own key routes spend to the provider rather than removing it, so exposure to usage driven cost is high.
Additional watchouts
Model token cost is the real expense and scales with usage and subagent delegation; credit purchases carry a 5% processing fee; cloud compute is billed separately.
Overage / add-ons
There is no overage in the subscription sense for the core tool. When prepaid Kilo Pass credits run low, users top them up or bring their own provider key and pay the provider directly, and credits roll over rather than expiring.
Sales call required
No, self serve available
Free / trial
Free and open source for individuals (VS Code, JetBrains and CLI); free models, local models and BYOK need no Kilo inference plan; 14 day Enterprise trial for teams.
Lowest paid plan
Teams $15/user/month (Kilo Pass inference credits from $19/month are optional)
Commercial notes
Individuals and Teams buy self serve; Enterprise is sold through sales with SSO, SCIM, audit logs and SLAs.
Key ambiguities
Real monthly cost depends on inference usage and model choice rather than the platform plan.
Cancellation / refund
The core extension and command line tool are free and open source with nothing to cancel. Kilo Pass credits roll over rather than expiring monthly. The KiloClaw cloud subscription has no contract and a one day free trial, so it can be canceled by the user.
Missing data
Enterprise pricing is custom; cloud compute rates are usage based.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose specialized AI agents build, host,…
- AgentUI — Managed AI app builder for operations teams that generates and hosts…
- Aider — Open source, model agnostic terminal coding agent that edits your…
- Anthropic Claude Code — Anthropic's agentic coding system across terminal, desktop, IDE, web…
- AppFactor — Agentic platform that runs persistent agents across an enterprise…
Alternatives to Kilo Code
The closest documented capability profiles to Kilo Code among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Kiro13.0 / 14Matches Kilo Code across all 14 documented capabilities
- GitHub Copilot13.5 / 14Fuller documented coverage on Observability & Auditability
- Refact.ai12.5 / 14A lighter documented profile than Kilo Code
- Cognition13.0 / 14Fuller documented coverage on Observability & Auditability
- OpenAI Codex13.0 / 14Fuller documented coverage on Observability & Auditability
- Warp13.0 / 14Fuller documented coverage on Observability & Auditability
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded