Available Actions
langchain_chat_completion
Chat completion action using Langchain
Parameters
The AI model to use (e.g., {:anthropic, [model: "claude-3-sonnet-20240229"]} or %Jido.AI.Model{})
type: {:custom, Jido.AI.Model, :validate_model_opts, []}
The prompt to use for the response
type: {:custom, Jido.AI.Prompt, :validate_prompt_opts, []}
List of Jido.Action modules for function calling
type: {:list, :atom}
Number of retries for validation failures
0
type: integer
Temperature for response randomness
0.7
type: float
Maximum tokens in response
1000
type: integer
Top p sampling parameter
type: float
Stop sequences
type: {:list, :string}
Request timeout in milliseconds
60000
type: integer
Enable streaming responses
type: boolean
Frequency penalty parameter
type: float
Presence penalty parameter
type: float
Forces model to output valid JSON (OpenAI only)
type: boolean
Enable verbose logging
type: boolean