[BUG] ChatLocalAI model not working

[Shekhar-Gudepu]

Below is the flow I created for a simple agent that uses SERP API to generate responses.

But I am getting multiple errors when I ask simple questions to the agent . Below are the errors,

Question: Hi Error: Failed to parse. Text: "". Error: SyntaxError: Unexpected end of JSON input Error: Failed to parse. Text: "". Error: SyntaxError: Unexpected end of JSON input

Question: What is flowise Failed to parse. Text: "Error: Failed to parse. Text: "ERROR: The prompt size exceeds the context window size and cannot be processed.". Error: SyntaxError: Unexppected token E in JSON at position 0 ". Error: SyntaxError: Unexpected token E in JSON at position 0

Expected behavior I expect the agent to call the SERP API if necessary and return correct responses.

Setup: Installation: docker
Flowise Version: flowise@1.3.0 OS: Windows Browser: Chrome

Note: The flow works as expected when the ChatOpenAI chat model is used. gpt4all.bin model was downloaded from (https://gpt4all.io/models/ggml-gpt4all-j.bin)

[HenryHengZJ]

Open source models like gpt4all is not capable for agent, they only work for simpler prompt using LLMChain:

[gregoryca]

is there a way to identify how to use the right models ? looks like the current implementation of chatlocalai is a bit limited.