Skip to content

bug: Missing response parsing for N8N Ai Agent Responses #8433

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
+55 −2
Open

bug: Missing response parsing for N8N Ai Agent Responses #8433

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
247e672
ollama n8n json response parsing
Psanyi89 Oct 24, 2025
936dc9d
remove debugging console.log
Psanyi89 Oct 24, 2025
11f5ab4
change scope of _isThinking variable and it's encapsulation
Psanyi89 Oct 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
57 changes: 55 additions & 2 deletions core/llm/llms/Ollama.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,6 +114,18 @@ type OllamaErrorResponse = {
error: string;
};

type N8nChatReponse = {
type: string;
content?: string;
metadata: {
nodeId: string;
nodeName: string;
itemIndex: number;
runIndex: number;
timestamps: number;
};
};

type OllamaRawResponse =
| OllamaErrorResponse
| (OllamaBaseResponse & {
Expand All @@ -124,7 +136,8 @@ type OllamaChatResponse =
| OllamaErrorResponse
| (OllamaBaseResponse & {
message: OllamaChatMessage;
});
})
| N8nChatReponse;

interface OllamaTool {
type: "function";
Expand Down Expand Up @@ -427,12 +440,52 @@ class Ollama extends BaseLLM implements ModelInstaller {
body: JSON.stringify(chatOptions),
signal,
});

let _isThinking: boolean = false;
function GetIsThinking(): boolean {
return _isThinking;
}
function SetIsThinking(newValue: boolean): void {
if (_isThinking !== newValue) {
_isThinking = newValue;
}
}
function convertChatMessage(res: OllamaChatResponse): ChatMessage[] {
if ("error" in res) {
throw new Error(res.error);
}

if ("type" in res) {
const { content } = res;

if (content === "<think>") {
SetIsThinking(true);
}

if (GetIsThinking() && content) {
const thinkingMessage: ThinkingChatMessage = {
role: "thinking",
content: content,
};

if (thinkingMessage) {
if (content === "</think>") {
SetIsThinking(false);
Copy link
Collaborator

@RomneyDa RomneyDa Nov 4, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

note we generally use camelcase for methods

}
// When Streaming you can't have both thinking and content
return [thinkingMessage];
}
}

if (content) {
const chatMessage: ChatMessage = {
Copy link
Collaborator

@RomneyDa RomneyDa Nov 4, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

other thoughts:
Could n8n have tool calls at all or not really?
Could we detect thinking from e.g. type field? On first look I wasn't able to pinpoint where docs for the N8nChatReponse would be, could you link them?

role: "assistant",
content: content,
};
return [chatMessage];
}
return [];
}

const { role, content, thinking, tool_calls: toolCalls } = res.message;

if (role === "tool") {
Expand Down
Loading