litellm support multi oauth chatgpt?

[ababbau]

Switch account waste time

[ashish993]

LiteLLM does not natively support multiple simultaneous OAuth ChatGPT accounts in a single proxy config, but here is how to handle multiple accounts:

Option 1 — Multiple model entries with different credentials
Define separate model entries in your config.yaml, each with its own OAuth token or API key:

model_list:
  - model_name: chatgpt-account-1
    litellm_params:
      model: openai/gpt-4o
      api_key: sk-account1-xxx
  - model_name: chatgpt-account-2
    litellm_params:
      model: openai/gpt-4o
      api_key: sk-account2-xxx

Then route to the appropriate model by name in your requests.

Option 2 — Load balancing across accounts
If you want automatic round-robin or least-busy routing across both accounts:

model_list:
  - model_name: chatgpt-multi
    litellm_params:
      model: openai/gpt-4o
      api_key: sk-account1-xxx
  - model_name: chatgpt-multi
    litellm_params:
      model: openai/gpt-4o
      api_key: sk-account2-xxx

router_settings:
  routing_strategy: least-busy

LiteLLM will automatically distribute load across both API keys.

For OAuth specifically (not API key): LiteLLM's OAuth support is model-provider dependent. For ChatGPT Enterprise with OAuth, you need to set api_base to your enterprise endpoint and pass the bearer token via api_key: Bearer <token>.

[omni-front]

Could you clarify what you mean by "multi oauth chatgpt"? Are you asking if LiteLLM can manage multiple ChatGPT accounts with OAuth authentication simultaneously? Or are you referring to switching between different accounts using OAuth? It would be helpful to know the specific scenario you're dealing with.