Providers

Create and manage credential providers that inject API keys and tokens into OpenShell sandboxes as environment variables.

AI agents typically need credentials to access external services: an API key for the model provider, a token for GitHub or GitLab, and so on. OpenShell manages these credentials as first-class entities called providers.

A provider is a named credential bundle. When you attach a provider to a sandbox, the gateway injects its credentials into the sandbox as environment variables at startup. Credentials never touch the sandbox filesystem.

When the trailing command in openshell sandbox create is a recognized tool name (claude, codex, or opencode), the CLI auto-creates the required provider from your local environment if one does not already exist. You do not need to create the provider separately.

Create a provider

From local credentials

The fastest way to create a provider is to let the CLI discover credentials from your shell environment:

openshell provider create --name my-claude --type claude --from-existing

This reads ANTHROPIC_API_KEY or CLAUDE_API_KEY from your current environment and stores them in the provider.

With explicit credentials

Supply a credential value directly:

openshell provider create --name my-api --type generic --credential API_KEY=sk-abc123

Bare key form

Pass a key name without a value to read it from the environment variable of that name:

openshell provider create --name my-api --type generic --credential API_KEY

This looks up the current value of $API_KEY in your shell and stores it.

Attach providers to sandboxes

Pass one or more --provider flags when creating a sandbox:

openshell sandbox create --provider my-claude --provider my-github -- claude

Each --provider flag attaches one provider. The sandbox receives all credentials from every attached provider at runtime.

Providers cannot be added to a running sandbox. If you need to attach an additional provider, delete the sandbox and recreate it with all required providers specified.

Manage providers

1. List all providers

   openshell provider list
   
2. Inspect a provider

   openshell provider get my-claude
   
3. Update a provider's credentials

   openshell provider update my-claude --from-existing
   
4. Delete a provider

   openshell provider delete my-claude
   

Supported provider types

The following provider types are supported. The --from-existing flag reads the listed environment variables from your current shell.

Type	Environment variables injected	Typical use
claude	ANTHROPIC_API_KEY, CLAUDE_API_KEY	Claude Code, Anthropic API
codex	OPENAI_API_KEY	OpenAI Codex
generic	User-defined	Any service with custom credentials
github	GITHUB_TOKEN, GH_TOKEN	GitHub API, gh CLI
gitlab	GITLAB_TOKEN, GLAB_TOKEN, CI_JOB_TOKEN	GitLab API, glab CLI
nvidia	NVIDIA_API_KEY	NVIDIA API Catalog
openai	OPENAI_API_KEY	Any OpenAI-compatible endpoint
opencode	OPENCODE_API_KEY, OPENROUTER_API_KEY, OPENAI_API_KEY	opencode tool

Use the generic type for any service not listed above. You define the environment variable names and values yourself with --credential.

Security model

Credentials are stored in the gateway — not on disk in the sandbox. The gateway injects them as environment variables at sandbox startup. This means:

• A compromised sandbox process cannot read credentials by scanning the filesystem.
• Credentials are purged when the sandbox is deleted.
• Providers can only be attached at sandbox creation time, not after the fact.

Supported inference providers

The following providers have been tested with inference.local. Any provider that exposes an OpenAI-compatible API works with the openai type. Set --config OPENAI_BASE_URL to the provider’s base URL.

Provider	Name	Type	Base URL
NVIDIA API Catalog	nvidia-prod	nvidia	https://integrate.api.nvidia.com/v1
Anthropic	anthropic-prod	anthropic	https://api.anthropic.com
Baseten	baseten	openai	https://inference.baseten.co/v1
Bitdeer AI	bitdeer	openai	https://api-inference.bitdeer.ai/v1
Deepinfra	deepinfra	openai	https://api.deepinfra.com/v1/openai
Groq	groq	openai	https://api.groq.com/openai/v1
Ollama (local)	ollama	openai	http://host.openshell.internal:11434/v1
LM Studio (local)	lmstudio	openai	http://host.openshell.internal:1234/v1

To configure inference routing, refer to Configure Inference Routing.

Next steps

Policies

Control what the agent can access on the network and filesystem.

Community sandboxes

Use a pre-built environment from the community catalog.

Policy schema reference

Full field-by-field YAML definition for sandbox policies.

GitHub sandbox tutorial

End-to-end walkthrough combining a GitHub provider with a scoped policy.