This morning, Cursor announced that it’s acquiring Graphite, the code review startup behind tooling used at Shopify, Snowflake, and Figma. The goal: combine Cursor’s AI code editor with Graphite’s review platform to build what they’re calling an “end-to-end” development experience.
The timing validates something we’ve been focused on for a while. As AI accelerates code generation, review becomes the constraint. Cursor CEO Michael Truell said it directly:
“The way engineering teams review code is increasingly becoming a bottleneck to them moving even faster as AI has been deployed more broadly within engineering teams.”
He’s identifying the right problem – the question is how you solve it without introducing more drag.
Two Paths to the Same Destination
Cursor’s approach is acquisition and integration. Buy an existing tool, then stitch it into the existing product over time. Truell mentioned they’ll be working on “deeper integration” throughout 2026.
Kilo’s approach is different: build it.
Code Reviews shipped within the Kilo platform two weeks ago. It analyzes PRs automatically when they’re opened or updated and provides inline AI-powered feedback and edit suggestions, surfacing issues across performance, security, style, and test coverage.
With Code Reviews, you can choose any model (including 3 that are currently completely free), adjust strictness, select priorities, and give custom instructions.
There’s no integration roadmap because there’s nothing to integrate. It’s the same platform you use for coding, debugging, and deploying. Enable it in the Kilo Dashboard here.
Why Native Matters
When code review lives inside the same system as your agents and sessions, interesting things happen.
Context flows naturally. The AI that reviews your code understands how it was written. Issues flagged in review can inform your next coding session. When you’re ready to ship, deployment is one click away with Kilo Deploy.
That’s harder to achieve when you’re connecting two products built by different teams with different architectures. Not impossible, but harder – and with unneeded friction.
The Model Lock-in Issue
Here’s something worth considering: Cursor uses Anthropic’s Claude as its primary model family, and Graphite is also Anthropic-backed. That’s not a conspiracy, but it does raise a practical question about how open model selection stays as the products merge.
Kilo takes a different stance. You choose from 500+ models for any task. Code generation, code review, debugging – whatever you need to do throughout your entire development workflow. You can use Claude, GPT-5, Gemini, Mistral, MiniMax, or bring your own API key. You can even switch models mid-session if you want.
The best model for your codebase might not be the best model for someone else’s. That should be your call – and that should extend to every part of your workflow.
Pricing Stays Simple
Acquisitions have costs, and those costs flow somewhere. We don’t know the terms of the Cursor-Graphite deal, but Graphite raised $52M in March, which gives you a sense of the scale.
Kilo’s pricing model is straightforward: you pay per token at the rate set by model providers. Code Reviews, Cloud Agent, Managed Indexing – they’re all part of the platform. No hidden costs for features that should be standard in the age of agentic engineering.
When we add capabilities, we build them – the economics stay the same, and you never have to worry about how model or platform lock-in will affect your workflow in an environment where standards are changing by the day.
What You Get Today
Kilo’s Code Reviews are available now, and you can try with a free account to get:
Automatic PR analysis the moment you open or update a pull request
Structured feedback across performance, security, style, and test coverage
500+ model options so you pick the reviewer that fits your needs
Eliminated friction by using the same platform you use for everything else in your workflow
No waiting for integration timelines. No hoping that two products play nice together. Just a feature that works, and lets you build at Kilo Speed.
The Bigger Picture
The AI coding market is moving fast. Cursor’s acquisition strategy signals they’re building toward a more complete platform. So are we. So is everyone in this space.
The difference is in the architecture. Kilo is open: open source, open model selection, open pricing. That’s not a tagline; it’s how we build. When you’re not locked to a single model provider or a single vendor’s roadmap, you keep your options open as the market evolves, and never sacrifice speed at the whim of someone else's roadmap.
Code review is becoming essential infrastructure for AI-accelerated teams, and we plan to continue delivering the experience at Kilo Speed.
Agreed 💯