Artificial Intelligence

MiniMax-AI Releases OpenAgentCore: An Open-Source, Self-Hosted Implementation of the OpenAI Agents API

· GlobeWire Daily Editorial Desk · Source: github.com

2 min read

Key takeawaysMiniMax-AI has published OpenAgentCore, an open-source and self-hosted project designed to replicate the OpenAI Agents API specification. According to the public repository documentation, the project allows developers to run coding assistants such as Codex, Claude Code, or MiniMax Code on local infrastructure or custom environments using the official…
MiniMax-AI Releases OpenAgentCore: An Open-Source, Self-Hosted Implementation of the OpenAI Agents API

Image: GitHub · Photo source · Resized and letterboxed to 16:9; no AI generation

Overview of OpenAgentCore

MiniMax-AI has published OpenAgentCore, an open-source and self-hosted project designed to replicate the OpenAI Agents API specification. According to the public repository documentation, the project allows developers to run coding assistants such as Codex, Claude Code, or MiniMax Code on local infrastructure or custom environments using the official OpenAI SDK or standard HTTP requests.

Core Features and Compatibility

  • OpenAI API Compatibility: The platform implements an API structure matching OpenAI's standards. Developers can configure existing tools by pointing the `OPENAI_BASE_URL` to their self-hosted instance while using standard project API keys.
  • Harness Flexibility: Sessions can natively execute supported coding harnesses and connect with user-configured model providers.
  • Persistent Workspaces and Compute Management: The platform incorporates sandbox options (such as microsandbox and E2B) designed to release idle compute resources while retaining workspace files on independent storage for later resumption.
  • Modular Architecture: Core components—including Sandbox Providers, workspace filesystems, execution harnesses, and model providers—interact through defined protocols, allowing individual parts to be replaced independently.

Getting Started and Deployment

According to the installation documentation, OpenAgentCore can be deployed on Linux or macOS environments utilizing Docker for Core, Web, and PostgreSQL components.

  • Setup Steps:
  1. Run the Docker installation for Core, Web, and PostgreSQL.
  2. Log into the Web interface using the Core key generated during installation.
  3. Configure the public address, establish a default model, and generate a Project API key.
  4. Add sandbox capacity using a local node or an integration like E2B.

Once configured, applications can interact with the environment by setting the appropriate base URL and API key variables within the official OpenAI SDK.

Practical Implications and Verification

OpenAgentCore offers teams looking for self-hosted agent orchestration an alternative compatible with standard tooling. Because execution states and workspace files are managed through defined protocols, developers can audit the architecture and adapt components to specific compliance or infrastructure requirements.

To verify the original announcement, review the project repository, setup guides, and licensing details directly on the official GitHub page.

Share this article

About the author

GlobeWire Daily Editorial Desk — AI-assisted news based on attributed published sources.

More to read