Home/Imbue
Imbue open AI agent tool ecosystem icon

Imbue

Imbue builds an open agent-tool ecosystem for running coding agents in parallel, owning local coding workspaces, reviewing agent changes, managing agent credentials, teaching repositories to operate, and filtering algorithmic feeds.

Visit website
Imbue product suite for parallel coding agents, review, and credential control

Product overview

What is Imbue?

Imbue is an AI research and software company building user-controlled tools for an open agent ecosystem. Its current catalog is a suite rather than one assistant: mngr manages many coding agents across local or remote environments; Sculptor is an extensible parallel coding workspace; Vet catches coding-agent mistakes; Latchkey gives local agents access to APIs without exposing credentials directly; Blueprint handles larger one-shot coding tasks; and Keystone helps a repository teach agents how to run it.

The suite also includes Bouncer, an open-source feed-filtering product for browsers and iOS. Across the catalog, Imbue emphasizes local control, open source, composability, and the ability to choose agents and infrastructure. The tools have different installation paths and maturity levels, so users should review each product's documentation and permissions separately.

How to Use Imbue Tools

  1. Choose the specific problem: parallel agents, coding workspace, review, credentials, repository setup, large coding tasks, or feed filtering.
  2. Open that product's official documentation and confirm its supported operating systems and interfaces.
  3. Install the CLI, editor extension, desktop application, browser extension, or mobile application as required.
  4. Configure the selected agent, compute environment, repository, or service credentials with minimum necessary access.
  5. Test on a non-critical workspace and inspect logs, diffs, network permissions, and generated changes.
  6. Keep human review and normal source-control safeguards in place before merging or deploying agent work.

Product Suite

  • mngr: Runs coding agents in parallel across a laptop, Docker, Modal, or SSH-accessible hosts.
  • Sculptor: Provides an owned, extensible workspace for parallel coding agents on macOS and Linux.
  • Vet: Reviews changes to catch mistakes made by coding agents.
  • Latchkey: Keeps credentials local while agents make standard authenticated API requests.
  • Blueprint: Supports larger one-shot coding tasks through CLI and VS Code.
  • Keystone: Helps repositories expose the knowledge agents need to run and work correctly.
  • Bouncer: Filters and reshapes algorithmic feeds across browsers and iOS.

Use Cases

  • Parallel issue resolution: Launch a separate coding agent for each independent issue or failing test.
  • Agent workspace management: Inspect, message, clone, pause, and organize agents from one CLI.
  • Safer API access: Allow local agents to call services without placing raw tokens in their context.
  • Agent code review: Detect problematic generated changes before they enter the main branch.
  • Repository onboarding: Give coding agents repeatable instructions for building, testing, and operating a project.
  • User-controlled feeds: Filter unwanted feed content with transparent, open-source software.

Pricing and Availability

Imbue does not publish one commercial plan for the entire suite. The products page marks several tools as open source, and mngr is described as free MIT-licensed software; users pay their own model inference and compute costs. Other products may have different distribution or commercial terms, so pricing should be checked on each product page.

Frequently Asked Questions

Is Imbue one AI agent?

No. It is a company and product ecosystem containing several agent, coding, security, and user-control tools.

Can mngr run different coding agents?

Yes. Its current product page lists Claude, Codex, OpenCode, and other shell-launchable agents across local and remote hosts.

Are all Imbue products open source?

The catalog marks many tools as open source, but users should verify the license and availability of the specific product they choose.

Back to product directory

Related products

OpenAI Codex is an AI coding product for delegating scoped repository tasks and reviewing the resulting changes.

Claude Code is Anthropic’s AI coding product for iterative repository work through supported development surfaces.

Gemini CLI is Google’s open-source AI agent for terminal workflows and local development tasks.

Newsletter

Keep up with useful AI products

Get a concise selection of new products, practical use cases, and builder updates.