Gemini LLM¶
Executes Large Language Model calls to Google’s Gemini models through the LiteLLM provider layer. The node selects the Gemini backend (provider="gemini") and supports multiple Gemini/Gemma model aliases via an internal fallback mapping to ensure compatibility with expected model names.
Usage¶
Use this node anywhere you need LLM-generated text or reasoning powered by Google Gemini within a workflow. Select a Gemini or Gemma model, provide your prompt/messages, and connect its output to downstream logic, parsing, or display nodes. It is typically paired with input nodes (for dynamic prompts) and output or inspection nodes (to capture responses).
Inputs¶
| Field | Required | Type | Description | Example |
|---|---|---|---|---|
| model | False | Not specified | Model identifier to use. If a plain name is provided, the node maps it to a provider-qualified name when needed. Supported aliases include: gemini-2.5-flash, gemini-2.5-pro, gemini-2.0-flash, gemini-2.0-flash-001, gemini-2.0-flash-lite-001, gemini-2.0-flash-lite, gemma-3-1b-it, gemma-3-4b-it, gemma-3-12b-it, gemma-3n-e4b-it, gemma-3n-e2b-it, gemini-flash-latest, gemini-flash-lite-latest, gemini-pro-latest, gemini-2.5-flash-lite. | gemini-2.5-flash |
| prompt_or_messages | False | Not specified | The text prompt or chat messages to send to the model. Exact structure depends on the base LLM interface. | Write a short haiku about the ocean. |
| generation_parameters | False | Not specified | Optional tuning parameters such as temperature, max tokens, and top_p. Specific names and ranges depend on the underlying base node. | {"temperature": 0.7, "max_tokens": 256} |
Outputs¶
| Field | Type | Description | Example |
|---|---|---|---|
| response | Not specified | The model’s response payload (usually text; exact structure depends on the base LiteLLM node). | The ocean whispers blue horizons stretch afar secrets in the deep |
Important Notes¶
- This node is backed by LiteLLM with provider set to "gemini".
- It supports a fallback model mapping so you can use simpler names (e.g., "gemini-2.5-flash") which are internally mapped to provider-qualified forms (e.g., "gemini/gemini-2.5-flash").
- You must configure valid credentials for the Gemini provider in your environment or platform settings. Never hardcode secrets; use secure configuration (e.g.,
). - This node is counted under LLM usage limits in certain plan tiers. If your workflow reports limit errors, review your LLM node quota.
- Exact input/output shapes (prompt/messages format, return structure) are defined in the shared LiteLLM base implementation and may vary by platform version.
Troubleshooting¶
- Model not recognized: Ensure the model name is one of the supported aliases or the fully-qualified provider form (e.g., "gemini/gemini-2.5-flash").
- Authentication error: Verify that a valid Gemini API key is configured in your environment (e.g.,
) and that your account has access to the chosen model. - Empty or truncated output: Increase max tokens (if supported) or reduce prompt size. Check any rate/usage limits.
- Rate limit or quota errors: Reduce request frequency, choose a lighter model, or upgrade your plan to increase LLM quotas.
- Unexpected response structure: Confirm downstream nodes expect the correct format and that you are using compatible parameters supported by the base LiteLLM interface.