OpenRouter Launches Workspaces for Granular AI Project Management and Team Control

TL;DR

• OpenRouter has introduced Workspaces, enabling users to segment AI projects into distinct environments.
• Each workspace offers independent control over critical settings like API keys, routing defaults, and Guardrails.
• This feature is designed for developers managing multiple projects, enterprises with diverse teams, and agents needing separate staging/production environments.
• Existing OpenRouter setups are automatically housed within a 'Default workspace', maintaining continuity for current users.

OpenRouter has officially rolled out Workspaces, a new feature designed to streamline the organization and management of AI projects. This update allows users to create separate, isolated environments for distinct projects, teams, or even different stages of agent development, such as staging and production. Workspaces bring a new level of organizational flexibility and control, addressing the growing complexity faced by developers and enterprises leveraging OpenRouter's extensive model ecosystem. Users can learn more about this significant update on the Introducing Workspaces Announcement page, or browse all OpenRouter Announcements.

Central to the Workspaces concept is the ability to configure independent settings for each environment. This includes managing API keys scoped specifically to a workspace, defining Guardrails that set policies for keys and members, and implementing BYOK (Bring Your Own Provider Keys) either per workspace or shared. Furthermore, users can optimize Routing to prioritize cost, latency, throughput, or tool-calling quality. Other configurable elements include Presets for prompts and model configurations, Plugins for default API request behavior, Observability for integrations, and Members for precise team access control. A detailed breakdown of these capabilities is available in the Workspaces Feature Guide, with specific information on policies accessible at Workspace Guardrails.

While Workspaces offer granular control, some account-level settings remain shared across all environments. These global settings encompass Activity & Logs, Credits & Billing, Organization management, Management Keys, and Privacy, acting as overarching governance. Notably, workspaces inherit account-level data policies and allowed providers/models, meaning they can only set more restrictive guardrails, not override global policies. Organization admins possess comprehensive control, managing everything within every workspace, including creation, deletion, and member access, in addition to account-level aspects like billing. New workspaces can be easily created through the dashboard UI or programmatically using the management API via OpenRouter Workspaces.

Summary

• OpenRouter's Workspaces simplify project organization by segmenting AI initiatives into isolated environments.
• Each workspace provides dedicated controls for API keys, security guardrails, routing logic, and team access.
• This feature significantly enhances flexibility and management for diverse use cases, from individual developers to large enterprise teams.
• Workspaces operate within account-level policies, ensuring consistent governance while allowing for project-specific customization and greater control.

Source: Introducing Workspaces | OpenRouter