[Error - Langchain Integration - AI Tool Agent] Received tool input did not match expected schema at DynamicStructuredTool.call

Artur_Ferreira_Padua · October 2, 2024, 7:20pm

Describe the problem/error/question

I am geting a generic error every time I use a specific tool agent. It seems that the error is caused by an input on a specific tool, but I have already tried changing and better defining the input. The weird thing is that the error happens on random times and not in all executions.

What is the error message (if any)?

{
“errorMessage”: “Internal error”,
“errorDetails”: {},
“n8nDetails”: {
“n8nVersion”: “1.53.2 (Cloud)”,
“binaryDataMode”: “filesystem”,
“stackTrace”: [
“Error: Received tool input did not match expected schema”,
" at DynamicStructuredTool.call (/usr/local/lib/node_modules/n8n/node_modules/@langchain/core/dist/tools/index.cjs:88:19)“,
" at processTicksAndRejections (node:internal/process/task_queues:95:5)”,
" at /usr/local/lib/node_modules/n8n/node_modules/langchain/dist/agents/executor.cjs:479:27",
" at async Promise.all (index 0)“,
" at AgentExecutor._call (/usr/local/lib/node_modules/n8n/node_modules/langchain/dist/agents/executor.cjs:471:30)”,
" at AgentExecutor.invoke (/usr/local/lib/node_modules/n8n/node_modules/langchain/dist/chains/base.cjs:58:28)“,
" at Object.toolsAgentExecute (/usr/local/lib/node_modules/n8n/node_modules/@n8n/n8n-nodes-langchain/dist/nodes/agents/Agent/agents/ToolsAgent/execute.js:144:30)”,
" at Object.execute (/usr/local/lib/node_modules/n8n/node_modules/@n8n/n8n-nodes-langchain/dist/nodes/agents/Agent/Agent.node.js:349:20)“,
" at Workflow.runNode (/usr/local/lib/node_modules/n8n/node_modules/n8n-workflow/dist/Workflow.js:728:19)”,
" at /usr/local/lib/node_modules/n8n/node_modules/n8n-core/dist/WorkflowExecute.js:673:51",
" at /usr/local/lib/node_modules/n8n/node_modules/n8n-core/dist/WorkflowExecute.js:1086:20"
]
}
}

Information on your n8n setup

n8n version: 1.53.2
**Running n8n via n8n cloud
Operating system: windows

n8n · October 2, 2024, 7:20pm

It looks like your topic is missing some important information. Could you provide the following if applicable.

n8n version:
Database (default: SQLite):
n8n EXECUTIONS_PROCESS setting (default: own, main):
Running n8n via (Docker, npm, n8n cloud, desktop app):
Operating system:

Artur_Ferreira_Padua · October 2, 2024, 7:33pm

n8n version: 1.53.2
Database: default: SQLite
n8n EXECUTIONS_PROCESS setting: default: own, main
Running n8n via: n8n cloud
Operating system: Windows

ria · October 4, 2024, 8:47am

Hi @Artur_Ferreira_Padua

Thanks for posting here and welcome to the community!

Can you please update your n8n version to the latest ? I think this is an old bug, we fixed.

Artur_Ferreira_Padua · October 4, 2024, 2:06pm

Hi @ria ,

Thanks for the reply.
We have tested on the new version and the error keeps hapenning!

If you need to have a deeper look, here are the worflow IDs and executions we’re ecountering the error:

Workflow: at8lCsgaqn5jK1gH - Execution ID: 27430
Workflow: XBTuRl5hPMzRWkAe - Execution ID: 27432

ria · October 8, 2024, 1:13pm

Hi @Artur_Ferreira_Padua

Thanks for the update!

We do not have access to your workspace, therefore could you please share the workflow that you are getting this error, so we can try to reproduce it? (Ideally with some dummy data pinned)

Thanks!

Wayne · October 23, 2024, 12:55pm

@Artur_Ferreira_Padua, did you resolve this? I’m also getting this issue

ria · October 23, 2024, 1:12pm

Hi @Wayne @Artur_Ferreira_Padua

We released a possibly related fix with 1.64.0

github.com/n8n-io/n8n

fix(OpenAI Node): Fix tool parameter parsing issue

n8n-io:master ← n8n-io:ai-385-community-issue-http-tool-does-not-work-correctly-with

opened 10:57AM - 10 Oct 24 UTC

burivuhster

+1 -1

## Summary This PR adds a fix to the OpenAI Node's tool parameter parsing log…ic. It should improve reliability of the tool calling. ## Related Linear tickets, Github issues, and Community forum posts https://linear.app/n8n/issue/AI-385/community-issue-http-tool-does-not-work-correctly-with-openai-node https://community.n8n.io/t/issue-with-http-request-tool/56719 ## Review / Merge checklist - [x] PR title and summary are descriptive. ([conventions](../blob/master/.github/pull_request_title_conventions.md))  - [ ] [Docs updated](https://github.com/n8n-io/n8n-docs) or follow-up ticket created. - [ ] Tests included.  - [ ] PR Labeled with `release/backport` (if the PR is an urgent fix that needs to be backported)

Can you please update and see if that resolved it?

Wayne · October 23, 2024, 2:47pm

@ria I’ve just updated and now all my tools are returning undefined

Wayne · October 23, 2024, 2:55pm

Yep all my tools are completely broken after the update.

The issues seem to be when the endpoint requires authentication.

ria · October 23, 2024, 3:12pm

Hi @Wayne

Can you share your workflow for reproducing this? I will raise this with our AI engineers.

Thanks!

Wayne · October 23, 2024, 3:13pm

Is there a way I can share it privately ?

ria · October 23, 2024, 3:20pm

We just need to see how you’ve configured it and which nodes you’re using etc.
It does not share your credentials by posting here. And you can modify the JSON as well to remove sensitive information.

Tip for sharing your workflow in the forum

Pasting your n8n workflow

Ensure to copy your n8n workflow and paste it in the code block, that is in between the pairs of triple backticks, which also could be achieved by clicking </> (preformatted text) in the editor and pasting in your workflow.

```
<your workflow>
```

Make sure that you’ve removed any sensitive information from your workflow and include dummy data or pinned data as much as you can!

But if it’s more comfortable for you, you can also DM me.

Wayne · October 23, 2024, 3:24pm

I shall DM, there lots of work gone into this one and it’s a client project so not for public consumption

ria · October 23, 2024, 4:04pm

Hi @Wayne @Artur_Ferreira_Padua

Sorry guys I just saw this now, but we actually have another fix (in particular for the HTTP tools with predefined credential types) - this should help you @Wayne. You can follow it up here:

github.com/n8n-io/n8n

fix(HTTP Request Tool Node): Fix the undefined response issue when authentication is enabled

n8n-io:master ← n8n-io:ai-407-http-tool-returning-response-undefined

opened 10:25AM - 22 Oct 24 UTC

burivuhster

+183 -108

## Summary This PR fixes an undefined response from HTTP Tool node in case au…thentication is enabled. Since this PR was merged (https://github.com/n8n-io/n8n/pull/11219) HTTP tool node requires a full response. But the underlying implementation for the request may differ, depending on authentication setting, and so the name of the option to get the full response. HTTP Tool Node now uses `helpers.httpRequestWithAuthentication` instead of deprecated `helpers.requestWithAuthentication`, so now it should respect the `returnFullResponse` option. ## Related Linear tickets, Github issues, and Community forum posts - https://github.com/n8n-io/n8n/pull/11219 - https://linear.app/n8n/issue/AI-407/http-tool-returning-response-undefined - https://community.n8n.io/t/http-tool-request-error/58998 ## Review / Merge checklist - [ ] PR title and summary are descriptive. ([conventions](../blob/master/.github/pull_request_title_conventions.md))  - [ ] [Docs updated](https://github.com/n8n-io/n8n-docs) or follow-up ticket created. - [ ] Tests included.  - [ ] PR Labeled with `release/backport` (if the PR is an urgent fix that needs to be backported)

But that’s actually a different issue than the original forum post by @Artur_Ferreira_Padua - Would be good to hear some feedback from you as well.

Thanks!

Wayne · October 23, 2024, 4:09pm

Thanks Ria, is this under release 1.64.1? There are no release notes as yet.

ria · October 24, 2024, 7:53am

Hi @Wayne

This should be released today and will then be in 1.64.2

Artur_Ferreira_Padua · October 24, 2024, 5:20pm

Hey ria!
Tks for reaching out.
Unfortunately the fix you sent won’t be helpfull since I am using mostly VertexAI and Anthopic as LLM Providers. By the way, we’ve this week updated to the version 1.64.1 and the bug keeps happening.

Watching the log I have realized a behaviour pattern even when JSON Schemas and LLMs vary. And the behaviour is that on the first run, the Agent nail the job defining a correct json schema with the property data structured, but the agent runs the LLM one more time and then come the error. The autofixing output parser also hasn’t beem able to fix. I’ll provide as much data as possible in the examples and if you need something more I can DM.

Here is the workflow:

Here is the json the agent structured on the first run:
{
“action”: “Final Answer”,
“action_input”: “{\n "__structured__output": {\n "__structured__output__array": [\n {\n "name": "João da Silva",\n "role": "Titular",\n "maritalstatus": "Casado",\n "healthplan": "ALICE CONFORTO",\n "dentalplan": "",\n "coparticipation": "false",\n "phone": "1199709875",\n "email": "[email protected]",\n "companyrole": "",\n "dependent_of": "",\n "relationship": ""\n },\n {\n "name": "Maria fatima",\n "role": "Dependente",\n "maritalstatus": "Casado",\n "healthplan": "ALICE CONFORTO",\n "dentalplan": "",\n "coparticipation": "false",\n "phone": "",\n "email": "",\n "companyrole": "",\n "dependent_of": "João da Silva",\n "relationship": ""\n }\n ]\n }\n}”
}

Here is the result after the secund run and using the autofix output parser. Btw, I’ve already tried not using the autofix.
[
{
“output”: [
{
},
{
}
]
}
]

The unwanted behavior is either a empty json as above or this generic error message that I got in the log of the second run on the LLM:
"". Error: SyntaxError: Unexpected token '`', "json
{
“… is not valid JSON”

Could you please help us asap? This error is making out application very unstable and as I’ve researched it looks like that is an instruction on how the AI Agent is parsing the json that seems native to the n8n.

Antoni_Camilo · November 13, 2024, 4:27pm

I solved it by replacing Chat-GPT with Groq

system · February 11, 2025, 4:27pm

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.