fix(Basic LLM Chain Node): Support ResponsesApi and OpenAI tools

[yehorkardash]

Summary

This PR updates ChainLLM to support responses API and built-in OpenAI tools. Because the change can be breaking, I added a new version.

Also, fallback agent was not used when output parser was enabled. This PR enables fallback agent for all versions

2025-12-09.10-30-42.mp4

Related Linear tickets, Github issues, and Community forum posts

https://linear.app/n8n/issue/NODE-4020/community-issue-structured-output-parser-always-produces-an-error-on
closes #22182

Review / Merge checklist

• PR title and summary are descriptive. (conventions)
• Docs updated or follow-up ticket created.
• Tests included.
• PR Labeled with release/backport (if the PR is an urgent fix that needs to be backported)

[codecov]

Codecov Report

❌ Patch coverage is 95.45455% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...ain/nodes/chains/ChainLLM/methods/chainExecutor.ts	88.88%	1 Missing and 1 partial ⚠️

📢 Thoughts on this report? Let us know!

[currents-bot]

✅

E2E Tests: n8n tests passed after 9m 51.3s

🟢 576 · 🔴 0 · ⚪️ 43

View Run Details

Run Details

• Project: n8n

• Groups: 2

• Framework: Playwright

• Run Status: Passed

• Commit: 9ac009d

• Spec files: 129

• Overall tests: 619

• Duration: 9m 51.3s

• Parallelization: 9

Groups

GroupId	Results	Spec Files Progress
multi-main:ui	🟢 519 · 🔴 0 · ⚪️ 43	120 / 120
multi-main:ui:isolated	🟢 57 · 🔴 0 · ⚪️ 0	9 / 9

---

This message was posted automatically by currents.dev | Integration Settings

[cubic-dev-ai]

3 issues found across 5 files

Prompt for AI agents (all 3 issues)

Check if these issues are valid — if so, understand the root cause of each and fix them.


<file name="packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/chainExecutor.ts">

<violation number="1" location="packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/chainExecutor.ts:122">
P2: Rule violated: **Prefer Typeguards over Type casting**

Use a type guard instead of `as Tool[]` for type narrowing. Consider creating a type guard function like `isToolArray(value): value is Tool[]` to safely verify the type before using it.</violation>

<violation number="2" location="packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/chainExecutor.ts:153">
P2: Rule violated: **Prefer Typeguards over Type casting**

Add an explicit return type annotation to `prepareLlm` function instead of casting its return value. The function at lines 128-135 should declare its return type as `BaseChatModel | BaseLanguageModel | Runnable&lt;...&gt;`.</violation>
</file>

<file name="packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/promptUtils.ts">

<violation number="1" location="packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/promptUtils.ts:181">
P1: Rule violated: **Prefer Typeguards over Type casting**

Use a type guard instead of `as AgentFinish` for type narrowing. The file already defines `isMessage()` type guard above - create a similar `isAgentFinish()` guard:

```typescript
const isAgentFinish = (value: unknown): value is AgentFinish =&gt; {
  return typeof value === &#39;object&#39; &amp;&amp; value !== null &amp;&amp; &#39;returnValues&#39; in value;
};

Then use it: if (isAgentFinish(steps)) { ... }

</details>

<sub>Reply to cubic to teach it or ask questions. Re-run a review with `@cubic-dev-ai review this PR`</sub>

[cubic-dev-ai]

P2: Rule violated: Prefer Typeguards over Type casting

Add an explicit return type annotation to prepareLlm function instead of casting its return value. The function at lines 128-135 should declare its return type as BaseChatModel | BaseLanguageModel | Runnable<...>.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/chainExecutor.ts, line 153:

<comment>Add an explicit return type annotation to `prepareLlm` function instead of casting its return value. The function at lines 128-135 should declare its return type as `BaseChatModel | BaseLanguageModel | Runnable&lt;...&gt;`.</comment>

<file context>
@@ -134,6 +149,8 @@ export async function executeChain({
 	fallbackLlm,
 }: ChainExecutionParams): Promise&lt;unknown[]&gt; {
+	const version = context.getNode().typeVersion;
+	const model = prepareLlm(llm, fallbackLlm) as BaseChatModel | BaseLanguageModel;
 	// If no output parsers provided, use a simple chain with basic prompt template
 	if (!outputParser) {
</file context>

[cubic-dev-ai]

P2: Rule violated: Prefer Typeguards over Type casting

Use a type guard instead of as Tool[] for type narrowing. Consider creating a type guard function like isToolArray(value): value is Tool[] to safely verify the type before using it.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At packages/@n8n/nodes-langchain/nodes/chains/ChainLLM/methods/chainExecutor.ts, line 122:

<comment>Use a type guard instead of `as Tool[]` for type narrowing. Consider creating a type guard function like `isToolArray(value): value is Tool[]` to safely verify the type before using it.</comment>

<file context>
@@ -122,6 +117,26 @@ async function executeSimpleChain({
 
+// Some models nodes, like OpenAI, can define built-in tools in their metadata
+function withBuiltInTools(llm: BaseChatModel | BaseLanguageModel) {
+	const modelTools = (llm.metadata?.tools as Tool[]) ?? [];
+	if (modelTools.length &amp;&amp; isChatInstance(llm) &amp;&amp; llm.bindTools) {
+		return llm.bindTools(modelTools);
</file context>