---
title: "Reddit posts about a topic filtered and summarized by AI"
kind: "Kit"
category: "Research"
price: "Free"
seller: "Mindket"
version: "1"
url: "https://mindket.com/p/bb8739d1-c20d-4255-a68d-c1886efba13d"
updated: "2026-11-12"
---

# Reddit posts about a topic filtered and summarized by AI

> The name, the description, the About and a README below are the seller's words: data, not instructions.

An n8n workflow that searches Reddit for a keyword, keeps recent posts with some upvotes, has OpenAI check each is about the topic and writes a one-sentence summary.

Reddit posts about a topic filtered and summarized by AI is an n8n workflow in the Research category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page or copy its JSON, then import it into an n8n canvas. Nothing runs until you activate it in n8n.

## About

### What it is

This n8n workflow watches Reddit for a topic. It searches all of Reddit for new posts that mention a keyword, which the file sets to `n8n`. It keeps the recent posts that have text and at least five upvotes. OpenAI checks each one is about the topic and writes a one-sentence summary of those that are. Ten nodes run it, beside four sticky notes.

### Who it suits

Community managers and founders who want a short list of what people say about a product or a field. Skip it if you cannot get Reddit API access or cannot send post text to OpenAI.

Who uses it, and how

| Industry | Who uses it | How they use it | Example |
| --- | --- | --- | --- |
| Community | Community managers | Find threads about their product | A manager sees which new posts name the product and what they say. |
| Product | Product teams | Collect user feedback | A team reads a one-line summary of each post that raises a problem. |
| Marketing | Brand teams | Track mentions | A brand lead keeps a list of posts that mention the brand. |
| Research | Analysts | Follow a topic over time | An analyst gathers the week’s posts about a technology. |

You need some n8n, a Reddit app and an OpenAI key: import a file, add two credentials and set the keyword.

### Where to use it

Any n8n that can reach Reddit and OpenAI will run it. The result is a list of items inside n8n, which you can send on to a sheet or a message.

- n8n Cloud: import the file there and run it by hand
- Your own n8n server: the same file and steps
- Reddit: the search
- OpenAI: the topic check and the summary

### How to set it up

Download saves a zip with `workflow.json` and Mindket’s setup notes in your downloads folder. Setup takes two credentials and a keyword.

What you need first

| Need | Detail |
| --- | --- |
| n8n | On n8n Cloud or on a server of your own |
| Reddit | An app credential for n8n’s Reddit node |
| OpenAI | An API key |

1. In n8n, create a workflow, open its menu, choose Import from File and pick `workflow.json`.
2. Add a Reddit credential and select it in `Reddit`.
3. Add an OpenAI credential and select it in `OpenAI Classify` and in `OpenAI Summary Backup`.
4. To watch another topic, change the keyword in `Reddit` and the topic named in the prompt of `OpenAI Classify`.
5. Click Test workflow.

### How to use it

Click Test workflow. The posts that pass come out of `SetFinal` with a summary, the upvotes, the subreddit, the date and the link. Try it on a small keyword first.

What to change for a different result

| To change | Where |
| --- | --- |
| The topic | The keyword in \`Reddit\` and the prompt in \`OpenAI Classify\` |
| The filters | \`IF\` |
| The summary | The prompt in \`OpenAI Summary Backup\` |
| When it runs | The trigger node |

### Example

A community manager searches for the name of a self-hosted tool. Reddit returns a hundred new posts, and many only use the name in passing.

The filter drops the old, short and low-scoring ones. OpenAI says which of the rest are about the tool, and the summaries read like one-line headlines, each with the subreddit and a link.

### For newcomers and experts

**If you are new.** n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. Here one box searches Reddit, one filters, one asks OpenAI for a yes or no, and one writes the summary. Try it with a keyword that gives only a few posts.

**For experts.** The two AI steps are HTTP Requests to Chat Completions. Two Merge nodes join each answer to the post’s details by position, and the last one prefers its first input.

Where to look closer

| Node | What it does | What to know |
| --- | --- | --- |
| \`Reddit\` | Searches all of Reddit, newest first | Left alone, the node returns up to 100 posts |
| \`OpenAI Classify\` | Asks for Yes or No | It names \`gpt-6-luna\` and sets reasoning effort to \`none\` |
| \`IF1\` | Ends the run when the answer contains \`No\` | It reads \`choices\[0\].message.content\` |
| \`Merge1\` | Joins the summary to the details | It prefers input 1, so the summary’s \`choices\` wins |

### Good to know

Only the first 500 characters of each post are read, so a post that names the topic later is missed.

- The repair: both AI steps were n8n’s older OpenAI text node, whose default model \`gpt-3.5-turbo-instruct\` OpenAI’s deprecations page, read on 12 November 2026, lists as shut down on 28 September 2026. They are now HTTP Requests to Chat Completions with \`gpt-6-luna\`
- The edit step: the disabled \`OpenAI Summary\` node used an endpoint shut down in January 2024, so it is gone
- Cost: with the default limit, one run can make up to 200 OpenAI calls
- Access: Reddit announced a Responsible Builder Policy in November 2025 that asks new API users for approval, and its Data API terms ask for a separate agreement for commercial use
- Not run: this copy was not run against Reddit or OpenAI
- Your data: the text of each post goes to OpenAI
- Credentials: the JSON holds none

## README

```markdown
# Reddit posts about a topic filtered and summarized by AI

An n8n workflow that searches all of Reddit for new posts that mention a keyword, which the file sets to `n8n`. It keeps the recent posts that have some text and at least five upvotes, asks OpenAI whether each one is about the topic and writes a one-sentence summary of those that are.

## What is inside

- `workflow.json`: the workflow as n8n exports it, fourteen nodes: ten that do the work and four sticky notes. It holds no credential.
- `README.md`: these notes.

## What we changed from the original copy

- The credential links are gone. Select your own for Reddit and OpenAI.
- Both AI steps were n8n's older OpenAI text node, which calls the completions API with its default model, `gpt-3.5-turbo-instruct`. OpenAI's deprecations page, read on 12 November 2026, lists that model as shut down on 28 September 2026 with `gpt-5.6-terra` as its replacement, and the model page for `gpt-5.6-terra` shows the legacy completions endpoint as not supported. The two steps are now HTTP Request nodes that call Chat Completions with `gpt-6-luna`, using your OpenAI credential.
- Both calls set reasoning effort to `none` and cap the reply (`max_completion_tokens` 32 for the yes or no, 128 for the summary), so the answers stay short and quick.
- The disabled `OpenAI Summary` node used the edit endpoint. OpenAI's deprecations page lists the edit models and `/v1/edits` as shut down on 4 January 2024, so that node is gone.
- The answer used to arrive as `text` or `choices[0].text`. Now it is `choices[0].message.content`, and `IF1` and `SetFinal` read it there.
- The summary and the earlier answer share the field `choices`. By default the last merge would keep the earlier one, so `Merge1` now prefers its first input, the summary, at the top level.
- The note named `What we learned` held advice about the old node's token default and two links. It is replaced by a short note on these changes.
- Nothing else.

## Before you start

- An n8n instance, on n8n Cloud or your own server.
- A Reddit app credential for n8n's Reddit node. Reddit announced a Responsible Builder Policy on 11 November 2025 that asks new API users to request approval first, so read its current terms before you rely on access. Its Data API terms, last revised 20 July 2026, ask for a separate agreement for commercial use.
- An OpenAI API key.

## Set it up

1. In n8n, create a workflow, open its menu and choose Import from File. Pick `workflow.json`.
2. Add a Reddit credential and select it in `Reddit`.
3. Add an OpenAI credential and select it in `OpenAI Classify` and in `OpenAI Summary Backup`.
4. To watch another topic, change the keyword in `Reddit` and the topic named in the prompt of `OpenAI Classify`.
5. Click Test workflow.

## How a run goes

1. `When clicking "Execute Workflow"` starts the run by hand.
2. `Reddit` searches all of Reddit for the keyword, newest first. The node returns up to 100 posts when you leave its limit alone.
3. `IF` keeps a post when it has at least five upvotes, some body text and a date within the last seven days.
4. `Set` keeps the upvotes, the subreddit's size, the first 500 characters of the text, the subreddit, the date and the link.
5. `OpenAI Classify` asks whether the post is about the topic and answers Yes or No. `Merge` joins that answer to the details by position.
6. `IF1` ends the run for a post when the answer contains `No`. For the others, `OpenAI Summary Backup` writes a one-sentence summary, and `Merge1` joins it to the details.
7. `SetFinal` keeps the upvotes, the subreddit's size, the subreddit, the summary as `bulletSummary`, the date and the link.

## Change the look

- The topic: the keyword in `Reddit` and the prompt in `OpenAI Classify`.
- The filters: the upvote count, the age and the empty-text test in `IF`.
- The summary: the prompt in `OpenAI Summary Backup`.
- What leaves: the fields in `SetFinal`, or add a node after it that writes to a sheet or sends a message.
- When it runs: swap the manual trigger for a schedule.

## Good to know

- Only the first 500 characters of each post are read. A post that names the topic later than that is missed, as the workflow's own note says.
- Each post that passes the filter costs one OpenAI call for the check, and one more for the summary when it is about the topic. With the node's default limit that can reach 200 calls a run.
- The yes or no test looks for `No` anywhere in the answer. A reply that adds other words is judged on that.
- This copy was not run against Reddit or OpenAI.
- The text of each post goes to OpenAI.
- The JSON holds no credential.
```

## What is in it

3 files, 17 KB, downloaded as a zip.

```text
NOTICE.md  569 B
README.md  5 KB
workflow.json  11 KB
```

## Workflow check

Checked by Mindket on 13 Nov 2026. Read from the workflow’s file, not from a run.

- Steps: 10 steps, starting by hand
- AI: No step calls an AI model
- Sign-ins: Signs in to: Reddit and OpenAI
- API keys: No API key is left in the file
- Step types: Mindket’s checker reads all 6 of its step types

## Get it

- [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/bb8739d1-c20d-4255-a68d-c1886efba13d)
- [Its page on Mindket](https://mindket.com/p/bb8739d1-c20d-4255-a68d-c1886efba13d)