Groq LLM

Sends a text prompt to Groq-hosted large language models and returns the model’s response. Supports a system prompt for behavior control, optional context inputs, temperature, and max token settings. Model options are fetched from the Salt LLM service and fall back to a built-in list if the service is unavailable.

Usage

Use this node when you need to generate or transform text using Groq-backed models (e.g., Llama, Qwen, Kimi). Select a model, provide a clear prompt, optionally include a system prompt to shape the assistant’s behavior, and set temperature/max_tokens to control style and length. You can pass in up to four optional context strings and reference them in your prompt with placeholders (e.g., {{input_1}}). Chain the output into downstream nodes that need the generated text.

Inputs

Field	Required	Type	Description	Example
model	True	STRING	The Groq model to use. The list is loaded from the Salt LLM service and falls back to a curated set if the service is not reachable.	llama-3.1-8b-instant
system_prompt	True	STRING	Instructions that set role, tone, and formatting rules for the assistant. Leave empty to skip. Multiline supported.	You are a concise technical assistant. Prefer bullet points and short code comments.
prompt	True	DYNAMIC_STRING	The main user prompt. You can reference optional inputs using placeholders like {{input_1}} ... {{input_4}} to inject external context.	Summarize the following notes in 5 bullets: {{input_1}}
temperature	True	FLOAT	Controls creativity and variability. Lower is more deterministic; higher is more diverse.	0.5
max_tokens	True	INT	Maximum number of tokens to generate. Set 0 for provider default.	1024
input_1	False	STRING	Optional context string that can be referenced in the prompt as {{input_1}}.	Meeting notes from 2025-09-15...
input_2	False	STRING	Optional context string that can be referenced in the prompt as {{input_2}}.	Customer feedback CSV excerpt...
input_3	False	STRING	Optional context string that can be referenced in the prompt as {{input_3}}.	Style guide: use active voice.
input_4	False	STRING	Optional context string that can be referenced in the prompt as {{input_4}}.	Audience: non-technical executives.

Outputs

Field	Type	Description	Example
Output	STRING	The text generated by the selected Groq model.	Here are five concise bullet points summarizing the notes...

Important Notes

• Model list and fallback: The node fetches available Groq models from the Salt LLM service and caches them for performance. If fetching fails, it uses a built-in fallback mapping (e.g., groq/llama-3.1-8b-instant).
• Prompt context placeholders: You can inject optional inputs into your prompt using {{input_1}} to {{input_4}}. Ensure the placeholders match the input names exactly.
• Max tokens and truncation: If the response hits the max_tokens limit, the output may be truncated by the provider.
• Temperature range: Valid range is 0.0 to 1.0; adjust to balance determinism vs. creativity.
• Latency considerations: Large models or long prompts can take time to respond. The node is designed to handle long-running requests.
• Default behavior: If not specified otherwise, the node applies a concise default system prompt to keep responses direct and focused.

Troubleshooting

• Model ID not found: If you see an error about an unknown model, reselect a model from the dropdown to refresh the list, or choose a known fallback option.
• Truncated responses: If outputs seem cut off, increase max_tokens or simplify the prompt.
• No placeholder substitution: Ensure your prompt uses valid placeholders (e.g., {{input_1}}) and that the corresponding inputs are connected/populated.
• Timeout or service error: Check network connectivity and service status. Retry the request or select a smaller model to reduce latency.
• Unexpected style or tone: Set or adjust the system_prompt to explicitly direct behavior (tone, format, and constraints).