# Kilo Code — Open source AI coding agent for VS Code, JetBrains, CLI and cloud, with 500+ models at provider rates.

> Source: The Agents Index — https://theagentsindex.com/kilo-code (structured, researched, re-verified)
> Facts last verified: 2026-09-02

Kilo Code is an MIT-licensed AI coding agent that runs in VS Code, JetBrains IDEs, the terminal, Slack and isolated cloud containers. It routes work to hundreds of hosted models, your own provider keys, or local models through Ollama and LM Studio.

| Fact | Value |
| --- | --- |
| Website | https://kilo.ai |
| Price | from $15/month |
| Free tier | Yes |

The VS Code extension, JetBrains plugin and CLI are free and open source with no card required; inference is separate and can be free models, local models or your own keys. A 14-day Enterprise trial sits on top of the free tier.

## What it does

**Writes code** · Yes — Authors and refactors code from a prompt across VS Code, JetBrains and the terminal.

> Kilo Code is an open source AI coding agent that runs inside VS Code, JetBrains IDEs, and your terminal. It writes, debugs, refactors, and reviews code, automates multi-step tasks across your repo
> — https://kilo.ai/ · Frequently asked questions, "What is Kilo Code?"

**Debugs and fixes** · Yes — Debug mode reads errors, traces the issue and proposes the fix.

> Debug Mode. Identify and fix bugs fast with an agent that reads errors, traces issues, and suggests fixes.
> — https://kilo.ai/ · "Built for agentic engineering" section, Debug Mode bullet

**Writes tests** · Yes — Test writing is named as part of the SDLC coverage Kilo sells to enterprises.

> Full SDLC — code review, refactoring, test writing, docs, and more
> — https://kilo.ai/enterprise · "Which Enterprise Challenge Is Yours?", Multi-Team & Subsidiary Orgs card

**Reviews code** · Yes — Kilo Code Reviewer analyses pull requests and comments in the PR.

> Kilo Code Reviewer analyzes your pull requests using AI models of your choice. It understands your codebase context, identifies issues, and provides actionable feedback directly in your PR.
> — https://kilo.ai/code-reviewer · FAQ, "How does Kilo Code Reviewer work?"

**Whole-codebase work** · Yes — Indexes the repository and runs multi-step repo-wide work such as migrations.

> Kilo automatically indexes your repository, pulls in the most relevant files and symbols for every prompt, and uses ripgrep-powered search, AST-aware tooling, and persistent memory
> — https://kilo.ai/ · Frequently asked questions, "Does Kilo understand my entire codebase?"

**Ships it** · Partly — Cloud Agents commit and push to a session branch and raise pull requests; the vendor names deployments only in a comparison claim, and documents no deploy pipeline.

> After every message, the agent commits and pushes your changes to the session branch on GitHub or GitLab.
> — https://kilo.ai/cloud · FAQ, "Is my work saved if the container shuts down?"

**Done-for-you service** · No — You run the software; the only people-delivered element is enterprise onboarding.

> High-touch onboarding: Speed Run embeds a Kilo engineer with your team
> — https://kilo.ai/enterprise · "Which Enterprise Challenge Is Yours?", Multi-Team & Subsidiary Orgs card

## How much it does unattended

**Runs autonomously** · Yes — Cloud Agents run long tasks unattended in isolated containers and open PRs.

> Cloud Agents run Kilo inside isolated cloud containers connected to your GitHub or GitLab repo, so long-running tasks like migrations, refactors, or autonomous PR creation keep working even when your laptop is closed.
> — https://kilo.ai/ · Frequently asked questions, "How do Cloud Agents work?"

**Multi-agent** · Yes — Orchestrator mode coordinates other modes; the CLI and JetBrains plugin run parallel agents in isolated worktrees.

> Start multiple specialized agents, monitor their progress side by side, and keep long-running coding tasks moving without opening extra tools.
> — https://kilo.ai/cli · "AI coding from your terminal", "Run parallel agents from one terminal"

**Agent permissions** · Partly — Auto-approve exists, but the OS-level sandbox that bounds it is experimental and Linux/macOS only.

> Agent sandboxing adds operating-system-level file system and network restrictions for safer auto-approve workflows. It is experimental, available on Linux and macOS first
> — https://kilo.ai/enterprise · FAQ, "How does agent sandboxing work?"

## Which models it runs on

**Claude** · Yes — Claude models are in the hosted catalogue and available via an Anthropic BYOK key.

> Kilo gives you access to 500+ AI models through one unified interface, including Claude Sonnet 5, Claude Opus 4.8, GPT-5.5, Gemini 3.1 Pro Preview, DeepSeek, Mistral, xAI Grok, Qwen, Llama
> — https://kilo.ai/ · Frequently asked questions, "Which AI models does Kilo Code support?"

**GPT** · Yes — GPT models are in the hosted catalogue and available via an OpenAI BYOK key.

> You can bring your own API keys for OpenAI, Anthropic, Google, Groq, and any OpenAI-compatible provider, and pay those providers directly at list prices.
> — https://kilo.ai/ · Frequently asked questions, "How does pricing work and can I bring my own API key?"

**Open models** · Yes — Open-weight hosted models plus fully local models through Ollama or LM Studio.

> You can also bring your own API keys for any provider, or run fully local models through Ollama or LM Studio when you need to keep code on your machine.
> — https://kilo.ai/ · Frequently asked questions, "Which AI models does Kilo Code support?"

**Model choice** · Yes — You pick the model; Auto Model routing is optional and never silently substitutes.

> Yes. Auto Model is a default routing option, not a lock-in. You can still pick a specific hosted model, use BYOK provider keys, or run local models with Ollama or LM Studio.
> — https://kilo.ai/auto-model · FAQ, "Can I still choose a model manually?"

## Where you use it

**In your editor** · Yes — VS Code and every VS Code-based editor, plus all JetBrains IDEs.

> Kilo Code runs in VS Code and every VS Code-based editor (including Cursor forks and Windsurf), in all JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, GoLand, Rider, RubyMine, CLion, and more)
> — https://kilo.ai/ · Frequently asked questions, "What IDEs and platforms does Kilo support?"

**On the command line** · Yes — Standalone CLI installed from npm, with sandboxed auto mode.

> Run kilo in any directory and start coding with AI assistance.
> — https://kilo.ai/cli · "Get started in minutes", step 3

**In your pipeline** · Yes — Reviews can fire automatically on PR creation, and Cloud Agent sessions can be started by webhook.

> Kilo Code Reviewer integrates with GitHub, GitLab, and Bitbucket (coming soon). You can trigger reviews automatically on PR creation or manually when needed.
> — https://kilo.ai/code-reviewer · FAQ, "How do I integrate it with my workflow?"

**In a browser** · Yes — Cloud Agents are driven from a browser dashboard, and there is a Slack surface.

> You can kick them off from the IDE, CLI, Slack, or a web dashboard, review their output as pull requests, and resume the exact same session locally with full context preserved.
> — https://kilo.ai/ · Frequently asked questions, "How do Cloud Agents work?"

## Whose machine it runs on

**Self-hosted** · Partly — The MIT clients can be forked and self-hosted and inference can be fully local, but Gateway, Cloud Agents and team administration are Kilo-hosted.

> Hosted services such as Kilo Gateway, Cloud Agents, and team administration add managed infrastructure and enterprise controls.
> — https://kilo.ai/ · Frequently asked questions, "Is Kilo Code really open source?"

**Open source** · Yes — VS Code extension, JetBrains plugin and CLI are MIT; the Gateway and Cloud backend are source available rather than MIT.

> Yes. Kilo's VS Code extension, JetBrains plugin, and CLI are open source under the MIT license. Kilo's Gateway and Cloud backend are published separately as source available, with security and abuse-protection code excluded.
> — https://kilo.ai/open · FAQ, "Is Kilo Code open source?"

## What it costs to run

**How it meters** · Yes — Three meters: per seat for the platform plan, provider token rates for inference, per second for cloud compute.

> Kilo separates platform access, AI inference, and cloud compute so you only pay for what you use and need.
> — https://kilo.ai/pricing · "Understand the three parts of pricing" intro

**Free tier** · Yes — The agent is free and needs no card; free and local models can carry the inference.

> No. You can download Kilo Code and start coding with AI immediately using free models, local models (Ollama, LM Studio), or your own API keys.
> — https://kilo.ai/pricing · FAQ, "Do I need a subscription to start coding with Kilo?"

**API access** · Partly — Cloud Agent sessions can be launched by HTTP webhook and usage data has documented OpenAPI endpoints; no public API for driving the agent itself is documented.

> Webhook triggers allow you to initiate Cloud Agent sessions via HTTP requests. You can reference request data in prompt templates using placeholders like {{body}}, {{method}}, and {{headers}}.
> — https://kilo.ai/cloud · FAQ, "How do webhook triggers work?"

**MCP server** · Yes — First-class MCP client with a marketplace and private org servers.

> Yes. Kilo is a first-class MCP client: you can connect any MCP server to expose custom tools, APIs, docs, or internal services to the agent.
> — https://kilo.ai/ · Frequently asked questions, "Does Kilo support MCP (Model Context Protocol)?"

**Bring your own key** · Yes — Own provider keys at list price, shared org keys on Teams, or fully local inference.

> Bring your own Anthropic, OpenAI, Google, Azure, AWS Bedrock, or other provider keys
> — https://kilo.ai/pricing · "Choose how you pay for AI inference", Auto Free/BYOK/local card

## Buying it for a team

**A company can buy it** · Yes — Teams is self-serve per seat; Enterprise is a custom contract through sales.

> Yes. Teams is self-serve, and Enterprise is a custom contract you arrange with sales.
> — https://kilo.ai/teams · FAQ, "Can we move from Teams to Enterprise later?"

**Seat model** · Partly — Flat $15 per user per month with Owner and Member roles; no seat floor or ceiling is published.

> The Teams platform plan is $15/user/month. AI model usage is billed separately at provider rates with no inference markup
> — https://kilo.ai/teams · FAQ, "How much does Kilo for Teams cost?"

**Pooled budget** · Yes — Teams share one Gateway balance and shared BYOK keys; Enterprise adds sub-org budgets and per-user limits.

> Teams adds centralized billing, a shared Kilo Gateway balance, shared BYOK keys, usage analytics and reporting
> — https://kilo.ai/teams · FAQ, "What's included in Kilo for Teams?"

**Admin controls** · Partly — Teams gets data-privacy controls on models and providers; allowlists, RBAC and policy enforcement are Enterprise.

> Yes. Enterprise admins can configure provider and model controls, shared BYOK routes, data policy filters, provider location filters, and organization-level policies so developers only see approved AI options
> — https://kilo.ai/enterprise · FAQ, "Can we restrict models and providers?"

**Audit log** · Partly — Audit logs are an Enterprise feature, absent from Free and Teams.

> Enterprise adds governance and security review: SSO/SAML, OIDC, SCIM provisioning, RBAC, audit logs, provider and model allowlists
> — https://kilo.ai/teams · FAQ, "What is the difference between Teams and Enterprise?"

**Single sign-on** · Partly — SAML, OIDC and SCIM provisioning are Enterprise-only.

> SSO/SAML, SCIM provisioning, OIDC, and RBAC keep access governed as teams and subsidiaries scale.
> — https://kilo.ai/enterprise · "Governance and security" section, Identity & Provisioning card

## What happens to your code

**Opt out of training** · Yes — Kilo states it never trains on your code and does not retain prompts for its own purposes; local models keep code on the machine.

> Kilo does not train on your code or prompts and does not retain them for its own purposes. Inference is sent to the provider your organization configures and remains subject to that provider policies.
> — https://kilo.ai/teams · FAQ, "Does Kilo train on our code?"

**Data residency** · Partly — EU residency is an Enterprise-scoped contractual configuration, not a switch, and depends on the providers chosen.

> Enterprise customers can scope a contractually backed EU data residency configuration. The exact commitment depends on the Kilo products, cloud compute, inference providers, endpoints, and data paths selected.
> — https://kilo.ai/eu · FAQ, "Can Kilo guarantee EU data residency?"

**Getting out** · Partly — Usage data has documented export endpoints and credit purchases can be refunded (the 5% fee is not); the site states nothing about credits expiring.

> The processing fee is non-refundable, including when the associated credit purchase is refunded.
> — https://kilo.ai/pricing · FAQ, "When does the 5% processing fee apply?"

**Certifications** · Partly — SOC 2 Type II is claimed with the report available through the Trust Center; no ISO 27001 or GDPR certification claim appears on the site.

> teams get enterprise controls including SSO, audit logs, SOC 2 Type II certification, and data-residency controls
> — https://kilo.ai/ · Frequently asked questions, "How does Kilo handle code privacy and data?"

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

Pages about this product that the vendor does not control. The label states how independent each one is.

| Source | Kind | Read |
| --- | --- | --- |
| bodegaone.ai | Independent source | 2026-08-18 |
| aiagentsquare.com | Independent source | 2026-08-18 |

## What changed

- 2026-09-02 — Capabilities · Commercial terms · Free tier · Price
- 2026-08-31 — Capabilities · Commercial terms · Free tier · Links · Price
- 2026-08-30 — Capabilities · Commercial terms · Lifecycle · Lifecycle evidence url · Links
- 2026-08-18 — Created · One-liner · Summary

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent Kilo Code, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.