Webhooks

[william-p]

Hi team,

For my own use case, I needed webhook support on questions and answers, so I went ahead and implemented a first iteration of the feature.

I'm not a Go expert, so I used Opus as a serious development aid to help me design and implement the logic. I tried to follow the existing architecture patterns in the codebase, using an event-driven approach where question and answer lifecycle events are dispatched to webhook handlers that asynchronously POST JSON payloads to configured endpoints.

Here's my fork: Work In Progress

What do you think? Would you be open to me submitting a pull request?

[LinkinStars]

I needed webhook support on questions and answers

Could you elaborate on the functionality of webhook? If it's just about notifying users of new questions or answers, the existing notification plugins might already support some of those features.

FYI: https://github.com/apache/answer-plugins/tree/main/notification-slack

[william-p]

In my case, I need it to trigger N8N workflow to synchronize Answer with a QDrant vector database to feed a RAG.

[william-p]

Anyway, IMHO, this is a standard feature for exposing a tool to external world.

I can propose an initial implementation, as I need it anyway. However, I understand this is a significant change and that there may be some hesitation.

[william-p]

(ignore PR event in discussion thread, sorry for the noise)

Do you think a plugin is a better approach?

Here are my implementations in core: https://github.com/william-p/answer/commits/featWebhooks/

I intended to split this into multiple pull requests ...

[LinkinStars]

(ignore PR event in discussion thread, sorry for the noise)

Do you think a plugin is a better approach?

Here are my implementations in core: https://github.com/william-p/answer/commits/featWebhooks/

I intended to split this into multiple pull requests ...

There's already a PR (#1510) that might help with your use case. Your goal is to sync Answer data to QDrant for RAG, right? PR #1510 adds semantic search and embedding capabilities directly into Answer's AI chat. This means you may no longer need the external webhook → N8N → QDrant flow — you could potentially achieve the same RAG functionality entirely within Answer.

[william-p]

Okay, thanks