cto.new

Overview

cto.new is a free AI-powered code agent built by Engine Labs that plans and generates production-ready code using frontier models like GPT-5 Codex, Claude Sonnet 4.5, and Gemini Pro. It breaks complex projects into manageable tasks, writes code, runs linters, and submits pull requests. The platform integrates with GitHub, Linear, Jira, and Slack, fitting directly into your existing workflow.

Key Features

Multi-Model AI (Alloys): Uses multiple frontier models (GPT, Claude, Gemini) with automatic routing and selection per task.
Intelligent Task Planning: Breaks down complex development projects into manageable steps with clear implementation plans before writing any code.
Automated Code Generation & PR Creation: Generates production-ready code asynchronously and submits pull requests you can review, even while you step away.
Auto Linting & Code Quality: Runs your project’s configured linting, formatting, and test commands and fixes issues before PR submission.
Deep Workflow Integrations: Connects natively with GitHub, GitLab, Linear, Jira, and Slack so you can assign tasks by tagging @cto in issues or adding labels.
Codebase Context Learning: Builds context about your repository structure and patterns across tasks, improving consistency within the project.
Isolated Sandbox Environments: Executes and tests AI-generated code in safe, isolated environments before submitting changes to your repository.
IDE Integration via MCP: Works as a background agent from Cursor IDE, letting you keep coding while cto.new handles separate tasks.

Pros

Completely Free Access to Frontier Models: Use GPT-5 Codex, Claude Sonnet 4.5, and Gemini Pro without paying subscription fees or providing API keys.
No Setup Friction: Start coding immediately with no credit card, API keys, or complex configuration required.
Fits Into Existing Workflows: Native integrations with GitHub, Linear, Jira, and Slack mean you don't need to change how your team works.
Asynchronous Operation: Assign a task and walk away. The agent works in the background and delivers a pull request when done.
Responsive Community & Support: Active Discord community with a development team that responds quickly to feedback and feature requests.

Cons

Fair Usage Limits: Free plan comes with rolling 24-hour and 7-day usage caps that may restrict heavy or production-scale workloads.
Initial Learning Curve: Getting accustomed to the agent's interface, task assignment workflow, and behavior patterns takes some adjustment.

Use Cases

Solo Developer Acceleration: You're a solo developer or freelancer working across multiple repos. Assign feature tickets to cto.new from GitHub Issues and receive tested pull requests while you focus on architecture and client work.
Bug Resolution with Observability Tools: Connect Sentry or Cloudflare Observability via MCPs. When a production bug hits, assign it to cto.new and the agent uses error context to plan a fix, write the code, and submit a PR without manual intervention.
Startup MVP Development: Your early-stage team needs to ship fast on a tight budget. Use cto.new to prototype features, generate boilerplate, and draft PRs at zero cost while keeping your runway intact.
Onboarding New Team Members: Point new developers at cto.new to get explanations of your codebase, architectural decisions, and project workflows, cutting ramp-up time significantly.
Technical Debt Cleanup: Assign technical debt issues and refactoring tasks to the agent. It analyzes your codebase, identifies improvement areas, and submits clean PRs that follow your existing code conventions.
Multi-Agent Development Workflow: Combine cto.new with other tools. Use Claude for writing specs, cto.new for implementing features, and GitHub Copilot for code review to create a high-output development pipeline.

Frequently Asked Questions

Is cto.new really free?

cto.new provides access to multiple frontier models, including OpenAI GPT models, Anthropic Claude models, and Google Gemini models. Its “Alloys” system can route or combine multiple models within a single task, allowing different models to handle planning, coding, debugging, and review inside one workflow.

What AI models does cto.new use?

Your code remains private within your workspace and connected repositories. The platform uses encrypted communication and secure sandbox environments. It may use anonymized usage data and system telemetry to improve performance and reliability. Your repository content is not publicly exposed and is only accessible within your authenticated environment.

What tools does cto.new integrate with?

cto.new integrates directly with common developer tools including GitHub, GitLab, Linear, Jira, and Slack. It also supports MCP integrations, allowing connections to external tools such as Sentry, Vercel, and other observability or workflow systems. Developers can assign work directly from issues or project management tools, and the agent will generate code changes and pull requests based on those tasks.

Can I run multiple tasks at the same time?

Yes. You can run multiple tasks in parallel across different repositories. For best results, avoid overlapping edits on the same files, as this may lead to merge conflicts or conflicting changes between tasks.

How does cto.new compare to Cursor, Claude Code, or Devin?

cto.new is an asynchronous AI coding agent focused on full task execution rather than in-editor assistance. It handles:

planning
coding
testing
pull request creation Unlike tools that assist inside the IDE, cto.new operates in the background and integrates directly with your development workflow tools. Performance is evaluated using internal real-world benchmarks focused on task completion, not external leaderboard claims.

Is cto.new suitable for enterprise use?

cto.new is widely used by individual developers, freelancers, and startups. For enterprise adoption, teams should evaluate:

SOC2 compliance status (Type I completed, ongoing maturity improvements)
security and data handling requirements
early-stage product maturity It can be used in enterprise environments, but suitability depends on internal compliance and risk tolerance.