Help: How to use the API call for user feedback?

[ErinLqq]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing? Tell me about your story.

I want to use APIs to like or dislike user feedback. Is there any official documentation? Are there any mandatory request parameters?

2. Additional context or comments

No response

[dosubot[bot]]

The API endpoint for submitting user feedback (like/dislike) is:

POST /messages/{message_id}/feedbacks [1]

Required headers:

• Authorization: Bearer <your_api_key>

Request body:

{
  "rating": "like",
  "user": "end_user_id"
}

Parameter	Required	Description
rating	Yes	"like", "dislike", or null (to remove feedback)
user	Yes	A string identifier for the end user
content	No	Optional text feedback

[2]

Successful response:

{
  "result": "success"
}

Key notes:

• The message_id in the URL is the ID of the assistant message you want to rate [3].
• Setting rating to null will delete existing feedback for that message.
• This works for Chat, Agent Chat, and Advanced Chat app types [1].

You can also retrieve all feedbacks via GET /app/feedbacks?page=1&limit=20 [1].

To reply, just mention @dosu.

---

Docs are dead. Just use Dosu.

[ErinLqq]

Is the request http://API server/messages/{message_id}/feedbacks required?

[dosubot[bot]]

Yes, the request format is correct. The full URL should be:

POST https://{your_api_server}/v1/messages/{message_id}/feedbacks

[1]

• For Dify Cloud: https://api.dify.ai/v1/messages/{message_id}/feedbacks [1]
• For self-hosted: https://{your_server}/v1/messages/{message_id}/feedbacks

Make sure to include the /v1 prefix — it's required for all Service API endpoints [2].

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.

[ErinLqq]

I still cannot correctly find the message ID because I am using Dify to connect with Feishu. How can I obtain the message ID from the Dify response?

[ErinLqq]

The logs I can see look like this, and it appears that the message_id is consistent. So, I obtained the start message_id, but the request returned a 404 error. Therefore, how should I accurately obtain the correct message_id for the streaming output?

[dosubot[bot]]

The message_id from your streaming response (9dc5c501-a342-4387-a539-429c46d73253) is correct — that's the right ID to use for the feedback API.

The 404 is most likely caused by a mismatched user parameter. When Dify looks up the message, it validates that the user in your feedback request matches the from_end_user_id of the original message [1]. If they don't match, Dify returns 404 even though the message exists.

To fix this: ensure the user value in your feedback request is the exact same user identifier that was used when the original chat message was sent from your Feishu integration.

POST /v1/messages/9dc5c501-a342-4387-a539-429c46d73253/feedbacks
{
  "rating": "like",
  "user": "same-user-id-used-when-sending-the-original-message"
}

Looking at your streaming data, the user ID used to create the message appears to be 60c3d4fb-a4b9-46d5-9098-09bbc09eeac4 (from sys.user_id in the workflow inputs). Try using that value as your user parameter.

Also make sure you're calling the feedback endpoint after receiving the message_end event — if called during streaming, the message may not be fully persisted yet [2].

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.

[ErinLqq]

Hello, I would like to ask how I can export a table of data for user feedback and all my conversation information from the backend.

[dosubot[bot]]

Dify provides a few ways to export this data:

User Feedback Export (Console API)

GET /console/api/apps/{app_id}/feedbacks/export?format=csv

Supported query parameters [1]:

• format: "csv" (default) or "json"
• rating: filter by "like" or "dislike"
• from_source: filter by "user" or "admin"
• has_comment: boolean
• start_date / end_date: date range filter (YYYY-MM-DD)

This export includes the message query, AI response, rating, and comments [2].

Conversation/Message History

There's no REST API endpoint for bulk conversation export. You have two options:

1. Paginated read via Service API — use GET /v1/messages?conversation_id={id}&user={user} to iterate through messages per conversation [3]. You'd need to script this yourself.

2. CLI command (backend access required) — Dify has an export-app-messages CLI command that outputs all messages in JSONL.GZ format, including feedback and retriever resources [4].

Note: The feedback export endpoint is part of the Console API (requires admin/console authentication), not the Service API (/v1). If you're only using the Service API key, you can still list feedbacks in a paginated way via GET /v1/app/feedbacks?page=1&limit=20 [3] but won't get a single-file export.

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.

[ErinLqq]

Is there a version now that supports using skills in workflows? For example, I download the skills from the SkillHub official website, upload them to Diffy, and then use them in workflows.

[dosubot[bot]]

Yes, Dify supports plugins that can be used in workflows. There's no "SkillHub" per se, but Dify has its own Plugin Marketplace where you can discover and install plugins directly.

How to use plugins in workflows:

1. Install plugins — Dify supports installing plugins from multiple sources [1]:

   • From the official Dify Marketplace (built into the platform)
   • From GitHub releases
   • From local package files (upload)

2. Use in workflows — Installed plugins (tools) can be used inside workflows through Agent Nodes. The Agent Node in a workflow can access any plugin tools you've installed and configured [2] [3].

Supported plugin types include: Models, Tools, Agent Strategies, Datasources, Triggers, and Extensions [4].

The Dify Marketplace is accessible from within the Dify platform itself (not an external "SkillHub" website). You can browse available plugins there, install them with one click, and then configure them in your workflow's Agent nodes.

If you're looking for specific tool plugins (like search, code execution, API integrations, etc.), check the marketplace from your Dify instance under the plugins section.

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.

[ErinLqq]

Hi The locally deployed Dify has scanned the following vulnerabilities. Are there any solutions? To which version should it be upgraded to resolve this vulnerability?

Vulnerability Overview
Dify is an open-source LLM application development platform released by LangGenius.
Version 1.9.1 of Dify contains a security vulnerability caused by improper permissions, which may lead to unauthorized access to system configuration data.