Imbue
Imbue · Coding
Imbue is an AI research company that builds software for running coding agents under your own control. Its two main tools are Sculptor, a free open-source coding agent workspace for parallel agents, and mngr, a Unix-style CLI for launching agents on your laptop, a Docker container, or a remote sandbox. The pitch is simple. Your agents, your code, your machine.
About Imbue
What Is Imbue
Imbue is an AI company founded in 2021 by Kanjun Qiu and Josh Albrecht. It started as Generally Intelligent, a research group working on agents that could reason and act, and later narrowed toward practical developer tools. The current product line is built around coding agents, meaning software that writes, edits, and runs code on your behalf while you stay in charge of what ships.
The idea behind Imbue is that the AI tools programmers use should stay accountable to the person using them. Closed platforms tend to lock you into their models, their servers, and their pricing. Imbue takes the opposite route with open-source, MIT-licensed tools you install and run yourself. You bring your own model plan and your own machine. Imbue doesn't store your repos or train on your code.
The biggest limit worth knowing up front? These are developer tools, not consumer apps. Sculptor runs only on Mac (Apple Silicon) and Linux. There's no Windows or mobile version yet, and using it means you're comfortable with Git, branches, and the command line. If that describes you, the trade-off is a lot of control for very little cost.
Getting Started
- Create an Imbue account on imbue.com if you want to follow updates, or skip straight to the GitHub repos for the tools.
- For Sculptor, download the desktop app for Mac (Apple Silicon) or Linux and connect an existing Claude or Claude Max plan.
- For mngr, install the CLI with one line:
curl -fsSL https://raw.githubusercontent.com/imbue-ai/mngr/main/scripts/install.sh | bash. - Start a workspace or launch an agent, then let each agent work in its own Git worktree, branch, and terminal.
- Review the diffs from one window, then
git fetch && git checkoutinto your normal repo to pull the changes.
Product Information
A quick look at Imbue's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Professional developers running several coding agents at once
- Engineers who distrust closed platforms
- Teams already paying for Claude
Tasks
- Running multiple coding agents in parallel
- Launching agents on remote compute
- Spec writing, mock generation, and TDD bug fixing
- Fixing broken pipelines
Scenarios
- A solo dev shipping a feature while two agents handle tests and docs in isolated branches.
- A team reviewing agent output before it ever touches the main branch.
- A cost-conscious developer who wants agent compute to shut down automatically when idle.
Key features
A workspace instead of a chat window
Most agent tools hand you a single chat window and hope you keep track. Sculptor gives you a workspace where every active agent is visible from one surface. Each one has its own worktree, branch, terminal, and diff view, so you review and merge without jumping between tools.
Isolated by default
Sculptor keeps each agent in a separate Git worktree with its own branch. That means an agent experimenting with a risky refactor can't quietly break the work another agent is doing, which is the failure mode that makes parallel agents stressful in the first place. Starting a fresh workspace takes one click. Cleanup is just as quick.
Free, local, and open source
Sculptor runs on your computer, not Imbue's servers. There's no subscription beyond whatever you pay your model provider, and the full source code sits on GitHub under a permissive MIT license that lets you read it, fork it, or ship your own version. Imbue's stated policy is that it doesn't store your repositories or train on your code.
mngr: agents in parallel, anywhere
mngr is a Unix-style CLI that launches a coding agent on your laptop, a Modal sandbox, or a Docker container with the same commands, so you rarely have to learn a new interface when you switch hosts. It works with Claude, Codex, OpenCode, or any agent you can start from a shell. There's no managed service in the middle, just SSH, Git, and tmux.
Plugins for your own agents
Both tools are meant to be extended. mngr supports plugins that add agent types, providers, and commands. Sculptor ships with skills for spec writing, mock generation, and TDD bug fixing, and those skills run as full agents that can read your codebase and spawn subagents.
Keeps out of your existing tools
Sculptor doesn't try to replace your editor or merge tool. Agents push to a local Git remote, and you pull changes into your normal repo with a standard fetch and checkout. Your workflow stays yours.
Pros and cons
Pros
- Open-source and MIT-licensed, so you can inspect and modify the tools you depend on.
- Sculptor runs parallel coding agents from a single workspace with real branch isolation.
- Works with existing Claude plans and other model providers instead of locking you to Imbue's stack.
- Remote runs auto-shutdown when an agent goes idle, which keeps compute costs low.
- mngr launches agents across local, Docker, and SSH hosts with one consistent set of commands.
Cons
- No Windows or mobile support, so Windows developers are left out for now.
- Sculptor still needs a paid Claude or Claude Max plan, so it isn't free end to end.
- The setup assumes comfort with Git, branches, and the command line, which rules out non-technical users.
- Sculptor is honest about being unfinished, so you may hit rough edges and should expect to file issues.
Frequently asked questions
Imbue is an AI company that builds open-source developer tools for running coding agents under your own control. Its main products are Sculptor, a desktop workspace for parallel agents, and mngr, a CLI for launching agents on local or remote hosts.
Related content
Explore related tools, skills, and articles for Imbue.
Imbue Alternatives
Forefront
Forefront · CodingForefront is a web platform for building with open-source AI. It lets you fine-tune leading open-source language models on your own data, evaluate how they perform, and run them through an API or export them to host yourself. Developers who want the convenience of a closed-source platform but insist on owning their models and data are the target audience here.
Startkit
StartKit.AI · CodingStartkit is a boilerplate for building AI SaaS and AI wrapper products. Think of it as an AI startup boilerplate with the boring parts already wired up: authentication, Stripe and Lemon Squeezy payments, usage limits, transactional email, and an AI API starter that talks to OpenAI, Anthropic, Groq, or Llama. You clone the repo, set your price, and start on the part of your product that people actually pay for. It's Next.js under React and Tailwind, so most of the boilerplate code already feels familiar.
Testim
Tricentis · CodingTestim is an AI-powered test automation platform for building and running end-to-end tests across web, mobile, and Salesforce applications. It leans on machine learning to keep tests stable when an interface changes, so teams spend less time fixing broken selectors. Not bad for an automated testing tool you can start using today. You create tests by recording actions in a browser, then optionally add JavaScript when you need more control. It's a solid pick for busy QA teams.