Lightning Ai Authentication

generated: '2026-07-19'
method: searched
source: https://lightning.ai/docs/platform/developers/cli
docs:
- https://lightning.ai/docs/platform/build/sandbox
- https://lightning.ai/docs/platform/developers/sdk/sandbox/python-sandbox-sdk
- https://lightning.ai/docs/platform/developers/sdk/sandbox/js-sandbox-sdk
- https://lightning.ai/docs/platform/developers/cli
note: >-
  Lightning AI publishes no OpenAPI definition, so this profile is captured from
  the published SDK and CLI authentication documentation rather than derived from
  securitySchemes. No OAuth 2.0 or OpenID Connect surface is documented for the
  platform API, so there is no scopes/ artifact for this provider.
summary:
  types:
  - apiKey
  api_key_in:
  - header
  - environment
  oauth2_flows: []
  openid_connect: false
  mutual_tls: false
schemes:
- name: LightningApiKey
  type: apiKey
  description: >-
    A Lightning API key authenticates the SDKs and the CLI against the platform
    control plane. Keys are created in the UI and can be scoped to a specific
    teamspace and role. The teamspace-scoped key is the recommended form - the
    key itself selects the project, so the JS SDK takes no teamspace argument.
  key_prefix: sk-lit-
  scoping: teamspace + role
  provisioning: Organization settings -> Members -> API keys -> Create API Key
  environment_variables:
  - name: LIGHTNING_SANDBOX_API_KEY
    role: Primary auth for Sandbox (preferred)
  - name: LIGHTNING_API_KEY
    role: Legacy / general platform fallback
  - name: LIGHTNING_USER_ID
    role: Paired with LIGHTNING_API_KEY for CLI environment auth
  - name: LIGHTNING_CLOUD_URL
    role: Optional control-plane base URL override (default https://lightning.ai)
  precedence: >-
    Explicit configure() values -> LIGHTNING_SANDBOX_API_KEY ->
    LIGHTNING_API_KEY (legacy). For the base URL, configure({ baseUrl }) ->
    LIGHTNING_CLOUD_URL -> production default.
  in_code:
  - language: python
    example: 'Sandbox.configure(api_key="sk-lit-...", base_url="https://lightning.ai")'
  - language: javascript
    example: 'Sandbox.configure({ apiKey: "sk-lit-...", baseUrl: "https://lightning.ai" })'
  caveats:
  - >-
    Python SDK reads credentials at import time - setting os.environ after
    import is ignored. The JS SDK reads credentials lazily on each request.
  - >-
    A personal login key cannot be used where a teamspace-scoped key is
    required; sandbox creation fails with an explicit "not your personal login
    key" error.
- name: InteractiveLogin
  type: interactive
  description: >-
    The CLI can authenticate interactively via a browser-based login, which
    stores credentials locally for subsequent commands.
  command: lightning login
- name: ModelApiKey
  type: apiKey
  description: >-
    Lightning Model APIs authenticate with the same Lightning API key, passed to
    the litai client or supplied to OpenAI- and Anthropic-compatible clients as
    the API key with https://lightning.ai as the base URL.
  obtain_at: https://lightning.ai/lightning-ai/model-apis/models?showApiKey=true
  in_code:
  - language: python
    example: LLM(model="openai/gpt-5", api_key="<LIGHTNING_API_KEY>")
  compatible_clients:
  - client: OpenAI Python SDK
  - client: Anthropic / Claude Code
    config: ANTHROPIC_BASE_URL=https://lightning.ai/ with ANTHROPIC_API_KEY set
      to the Lightning API key