Kodus Ai
AI Code Review with Full Control Over Model Choice and Costs.
Last verified:
What is Kodus Ai?
Kodus Ai — odus is an open-source AI code review platform built for engineering teams who ship AI-generated code. Its flagship agent, Kody, automatically reviews pull requests on GitHub, GitLab, Bitbucket, and Azure DevOps, catching bugs, security vulnerabilities, performance issues, and style violations before code reaches production. Unlike traditional code review tools, Kody understands your entire codebase—not just the diff—by using semantic LLM analysis backed by repository-wide code search.
Kody learns how your team works by analyzing your coding standards, architecture patterns, and workflows. You can define custom rules in natural language or choose from hundreds of pre-made rules to enforce your team's best practices. The tool integrates with Jira, Notion, and Linear to bring business context into reviews, and automatically converts unimplemented suggestions into issues to track technical debt.
Kodus is designed for software development teams, indie developers, students, and enterprises who want full control over their AI code review process. It is model-agnostic, supporting OpenAI (GPT-4o, ChatGPT), Anthropic (Claude), Google (Gemini), Llama, and any OpenAI-compatible endpoint. Teams can self-host Kodus or use the hosted version, with zero markup on LLM costs—you pay the model provider directly at their listed prices.
Kodus Ai pricing
Pricing model: Freemium
Kodus offers 3 tiers: (1) Community—Free for indie devs, students, and small teams with unlimited PRs using your own API key, unlimited users, up to 10 Kody Rules, up to 3 active plugins, Kody Learnings and Memory, and Discord Community Support; (2) Teams—$10 per developer per month (or $8/dev annually with 20% discount), includes hosted by Kodus, unlimited PRs with your own API key, unlimited users, unlimited Kody Rules, unlimited active plugins, priority queue for Kody Agents, Engineering Metrics/Cockpit, and Discord + Email Support, with 14-day free trial and no credit card required; (3) Enterprise—Custom pricing with self-hosted or hosted option, SSO, unlimited PRs using Kodus AI Tokens API key, RBAC + audit logs, SOC 2 compliance, up to 5 hours/month dedicated support with personalized onboarding, and Private Discord Channel. You pay Kodus for the platform license and your LLM provider for API usage with zero markup on tokens. 30-100% discounts available for startups, open-source projects, and teams in weaker currency countries.
Kodus Ai pros
- 100% open source with all code available on GitHub for inspection
- Model-agnostic: use Claude, GPT-4, Gemini, Llama, or any OpenAI-compatible endpoint
- Zero markup on LLM costs—you pay providers directly at their rates
- Deep codebase context understanding, not just diff analysis
- Automatic pull request reviews integrated directly into Git workflow
- Customizable rules in natural language plus hundreds of pre-made rules
- Supports GitHub, GitLab, Bitbucket, and Azure DevOps with under 5-minute setup
- Learns team workflows, coding standards, and architecture over time
- Integrates with Jira, Notion, and Linear for business context in reviews
- Technical debt tracking by converting unimplemented suggestions into issues
- Self-hosted or hosted by Kodus options for full control
- Engineering metrics and Cockpit dashboard for productivity insights
- No code stored on servers—all processing happens in real-time
- Does not train models on customer data for privacy and security
- Reduces code review time by 30-60% according to customer testimonials
- Fine-tune severity or disable specific rules per repo or folder
- Monorepo support with multi-repo and folder-level config inheritance
- Priority queue for Kody Agents on Teams and Enterprise plans
Kodus Ai cons
- Requires bringing your own API key and paying token costs separately
- Free tier limited to 10 Kody Rules and 3 active plugins
- Teams plan requires $10/dev/month plus token costs on top
- Setup requires OAuth or token configuration with Git provider
- May produce false positives until it learns team patterns
- Does not replace human review—only augments it as required step
- Enterprise plan requires sales contact for custom pricing
- No per-unit task cost visible, making single-task estimation difficult
Frequently asked questions about Kodus Ai
What is Kodus and what does it do?
Kodus is an open-source AI code review platform. Its first agent, Kody, is an AI-powered pull request reviewer that automatically reviews PRs, understands your entire codebase (not just the diff), follows your team's rules, and runs in the cloud or self-hosted on GitHub, GitLab, Bitbucket, or Azure DevOps.
Is Kodus really open source?
Yes, 100% open source. All source code is available on GitHub for anyone to inspect, use, and improve. It is licensed under AGPLv3.
Can I choose which AI model Kodus uses?
Yes, Kodus is model-agnostic. You can use Claude, GPT-4, Gemini, Llama, or any OpenAI-compatible endpoint. Run Kody with your own API keys and choose whichever model fits best.
Does Kodus charge markup on AI token costs?
No, Kodus charges zero markup on LLM costs. You pay the model provider directly at their listed price with no hidden multipliers or per-seat AI surcharges. Kodus revenue comes from the platform subscription, not from reselling tokens.
Does Kodus store my source code?
No, Kodus does not store your source code. All processing happens in real-time, and no part of your repository is saved on their servers.
Does Kodus train its models with customer data?
No, Kodus does not train its models with customer data. Your data is processed securely and is never used to improve or retrain their AI.
Which Git platforms does Kodus support?
Kodus integrates with GitHub, GitLab, Bitbucket, and Azure DevOps. It integrates at the pull request level—it reads diffs, posts inline comments, and respects your existing review workflows. Setup takes under 5 minutes.
How does Kodus handle false positives?
Kody learns from your team's patterns over time, so intentional stuff stops being flagged. You can also fine-tune severity or even disable specific rules per repo or folder. The goal is to stay helpful, not noisy.
Does Kodus work with monorepos?
Yes, monorepos were one of Kodus's first use cases. Kody supports multi-repo and folder-level configs with inheritance, so you can keep global rules for the org and still adapt to each package. It auto-syncs everything from GitHub/GitLab so it scales cleanly.
Who gets charged in a team—do I need to add everyone?
No, you decide who is included in the Kodus team and will only be charged for those users. You have full control over team management and billing.