Typesense icon

Typesense

Interact with the Typesense search platform

Actions50

Conversation → Create

AI-generated

Overview

This node interacts with the Typesense search platform, specifically supporting operations on various resources including Conversations. For the Conversation resource with the Create operation, it allows users to create a new conversation model configuration for a large language model (LLM). This is useful for managing and storing conversational AI models and their configurations, including historical conversation data, system prompts, and API keys. Practical applications include setting up and managing LLM-based chatbots or conversational agents with customizable parameters.

Use Case Examples

  1. Creating a new conversation model configuration for an LLM like GPT-4 or Claude-3.
  2. Storing conversation history in a Typesense collection for retrieval and analysis.
  3. Configuring system prompts and additional options such as TTL for message expiration.

Properties

Name Meaning
JSON Input Whether to provide the model configuration as raw JSON.
Model Configuration (JSON) Model configuration as raw JSON input, used if JSON Input is true.
Model Name Name of the LLM model to use, e.g., gpt-4 or claude-3.
Max Bytes Maximum number of bytes to send to the LLM in every API call.
History Collection Typesense collection that stores historical conversations.
Additional Options Additional optional parameters such as model ID, API key, account ID, system prompt, TTL for message expiration, and vLLM service URL.

Output

JSON

  • model_name - Name of the LLM model used in the conversation.
  • max_bytes - Maximum bytes sent to the LLM per API call.
  • history_collection - Typesense collection storing conversation history.
  • additionalOptions
    • id - Explicit model ID if provided.
    • api_key - API key used for LLM service authentication.
    • account_id - Account ID for the LLM service.
    • system_prompt - System prompt instructions for the LLM.
    • ttl - Time-to-live for messages before deletion.
    • vllm_url - URL of the vLLM service.

Dependencies

  • Typesense API service
  • LLM service API key for authentication

Troubleshooting

  • Error loading collections or fields may occur if the Typesense client is not properly configured or the API key is invalid.
  • If the resource or operation is not supported, the node throws an error indicating unsupported resource.
  • Ensure that required parameters like model_name, max_bytes, and history_collection are provided when JSON Input is false.
  • If JSON Input is true, ensure the JSON configuration is valid and correctly formatted to avoid parsing errors.

Discussion