Is your enhancement related to a problem? Please describe
Eclipse Che Dashboard has no built-in way for users to select an AI coding tool or manage API keys for AI providers. Users must manually configure DevWorkspace specs and create Kubernetes Secrets, which is error-prone and not discoverable.
Describe the solution you'd like
1. AI provider and tool definitions
Separate provider metadata (name, icon, description) from tool injection details (container image, binary, env var). Multiple tools can share one provider. Definitions are served to the frontend via the server config endpoint and can be overridden by administrators through the CheCluster CR.
2. AI provider selector on workspace creation
Add a selector on the Get Started page so users can pick an AI provider when creating a workspace. The selected provider ID is passed as a factory URL parameter and the tool is injected into the DevWorkspace at creation time.
3. API key management in User Preferences
Add an "AI Providers Keys" tab in User Preferences for managing API keys:
- List configured keys with provider name and status
- Add, update, or delete keys
Keys are stored as Kubernetes Secrets with auto-mount labels so the DevWorkspace Controller injects them as environment variables.
4. AI tool info on Workspace Details page
Show the currently injected AI tool in the Overview tab. Allow users to view available tools and switch to a different one (workspace must be stopped).
5. Backend API
- Serve provider and tool definitions via server config
- REST endpoints for listing, creating, and deleting API key secrets scoped to the user's namespace
Describe alternatives you've considered
- CheCluster CR only: Administrators configure AI tools entirely via CR, but users get no agency to select tools or manage their own API keys.
- Flat tool definitions: Keeping provider metadata on the tool definition duplicates data when multiple tools share a provider.
Is your enhancement related to a problem? Please describe
Eclipse Che Dashboard has no built-in way for users to select an AI coding tool or manage API keys for AI providers. Users must manually configure DevWorkspace specs and create Kubernetes Secrets, which is error-prone and not discoverable.
Describe the solution you'd like
1. AI provider and tool definitions
Separate provider metadata (name, icon, description) from tool injection details (container image, binary, env var). Multiple tools can share one provider. Definitions are served to the frontend via the server config endpoint and can be overridden by administrators through the CheCluster CR.
2. AI provider selector on workspace creation
Add a selector on the Get Started page so users can pick an AI provider when creating a workspace. The selected provider ID is passed as a factory URL parameter and the tool is injected into the DevWorkspace at creation time.
3. API key management in User Preferences
Add an "AI Providers Keys" tab in User Preferences for managing API keys:
Keys are stored as Kubernetes Secrets with auto-mount labels so the DevWorkspace Controller injects them as environment variables.
4. AI tool info on Workspace Details page
Show the currently injected AI tool in the Overview tab. Allow users to view available tools and switch to a different one (workspace must be stopped).
5. Backend API
Describe alternatives you've considered