Sculptor is the first coding agent environment that helps you embed software engineering best practices.
Sculptor runs your code in a sandbox, letting you test code safely, solve issues in parallel, and assign tasks to agents, working alongside any editor.
Use Sculptor to resolve bugs, write tests, add new features, improve docs, fix style issues, and make your code better—whether it was written by a human or an LLM.
Sculptor is an early research preview. If you’re excited to help shape its development, sign up as a tester! You’ll get early access, free use for a limited time, and rad Imbue swag.
How Sculptor works
Get suggestions for improving your code
When you connect Sculptor to your codebase, it checks for issues in your code, like missing tests, hardcoded variables, memory leaks, race conditions, and more.
Launch agents to fix the issues, and see the diff before applying.
Fix identified issues, in parallel
Kick off as many fixes as you want, all in parallel. Review proposed fixes and chat with agents to refine them.
When you accept and apply a fix, the code is synced to your local editor.
Turn your own expertise into custom checks that catch issues you care about
Define LLM prompts (like “make sure error messages are informative”) or shell commands (like your test suites or linters) to automatically flag and fix code according to your own specific practices.
Run code in isolated sandboxes to fix bugs until all your tests pass, so you can experiment with solutions without risking your local dev environment.
Sculptor integrates with common frameworks and linters like pylint, ruff, pyre, etc. so you get first-class support for your favorite testing tools.
Write high-quality new code
Write code in your favorite editor, or use Sculptor to generate new code.
Run your custom checks automatically on any new code, so you can detect and fix issues as you work.
Join us as a tester!
If Sculptor looks fun or useful to you, we’d love your help as a tester!
At Imbue, our goal is to democratize not just coding, but software engineering (here’s why). We’re exploring how to enrich LLM coding workflows with engineering best practices so that anyone can build maintainable systems without years of specialized training.
To make this vision a reality, we need your help! As a tester, you’ll get early access, free use for a limited time, and rad Imbue swag :)
p.s. If you want to nerd out with us…
…we’ll have another post soon about the design principles that went into Sculptor, and directions we’re really excited about, enabled by this agent environment architecture.
For example: how might we enable more expressivity and play (perhaps via “rewind”)? Is it useful for people to share custom checks with each other? And how could we automatically identify and fix issues in production to make maintenance easier?
Follow us on Twitter or sign up for our mailing list to get the update!