Voice of the Customer Research

The Reddit
Research Agent

Reddit aggressively blocks AI and crawlers, and you need voice-of-the-customer research. This agent pack operates within Reddit's Terms of Service to generate customer research from one of the richest sources of conversation online. No API necessary.

From the creator of the Ivy League course, Startup Studio
Get my research agent $10 · instant download
The Reddit Research Agent running in Claude Code, spawning a separate verifier that checks every quote
Why it exists

Reddit is the most honest focus group online.
It is also blocking every agent that tries to read it.

AI can't actually read Reddit

Chatbots can't fetch Reddit live; Claude Code literally blocks reddit.com at the fetch level and filters it out of web search. Ask one what Reddit thinks anyway and it quietly answers from stale training data, or hands you threads that were never posted. When you're betting a roadmap on the answer, invented evidence is worse than nothing.

This agent reads the real ones

You browse and save the threads that matter; the agent reads those exact files and writes the report. Every line traces back to a real post you can click and verify. Nothing automated ever touches Reddit, and you stay within its Terms of Service.

The product

A disciplined research analyst, as an agent pack

Install the agent in Claude Code, Codex, or Cursor and it runs a tested, repeatable process, pulling you in only where Reddit needs a human.

From the creator of the Ivy League course, Startup Studio.

Guided and fast

It runs; you step in only where needed

The agent drives the whole process and pulls you in only at the moments it needs you to grab a file, then carries on. The loop is tuned so keeping a human in it still moves quickly.

Knows where to look

Targets the right subreddits

It names the 3 to 5 communities worth mining for your topic, and which sort to use to surface enduring views and the objections people downvote.

Verified, not promised

A separate agent checks every quote

When the report is drafted, it automatically spawns a second agent with its own clean context to confirm each quote appears in the threads you saved. Anything it can't verify is fixed, dropped, or flagged. That is why the quotes hold up.

Finds the signal

Patterns, not anecdotes

It mines the high-upvote comments where real opinions live, and separates a theme that recurs across many threads from a one-off.

Honest about bias

Names the blind spots

It dates every quote, flags when evidence skews old, and calls out astroturfing, vocal minorities, and Reddit's own demographic skew.

Reddit ToS compliant

Built to keep you in bounds

A person collects the content and the agent analyzes it. Every page in the pipeline was loaded by a human in a normal browser session, so the whole workflow stays inside what Reddit allows. Compliant by design.

Within Reddit's TOS

Compliant by design

Reddit blocks bots and scrapers to keep the site valuable for its users. This agent pack respects that and works with you the human to stay in-bounds.

This is working within Reddit's limits, not around them.

What's inside the Reddit Research Agent pack
What you get

A complete agent pack,
not just a prompt

  • The researcher agent: scopes your topic, lists the threads to save, and writes the report.
  • A separate verifier agent: runs automatically to check every quote against your saved threads.
  • Helper scripts & guides: normalize your saved threads (plain Python that runs as-is, without installs or API keys), plus the collection and report-structure guides.
  • A complete example report: see exactly what good looks like before you run your own.
  • A clear single-user license: use it on as many of your own projects as you like.
Who it's for

Built for the scrappy early stage

When you need real customer signal fast and can't afford a research agency, or to be wrong.

Founders

Validate a problem in your customers' own words before you build the wrong thing.

Small businesses

Understand what your market actually complains about, and what makes them stay or leave.

Consultants & indie hackers

Run trustworthy voice-of-customer research for clients without a heavy toolchain.

Questions

Good to know

Does this scrape Reddit or break its rules?

No. Reddit aggressively blocks crawlers and agents, so this works the other way around. A refined, guided workflow lets you quickly save the threads that matter, and the agent analyzes only what you hand it. Nothing automated ever touches Reddit, and you stay inside its terms of service. The pack is built around that line on purpose.

If a human is in the loop, is it slow?

No. The agent runs the process and pulls you in only at the exact moments it needs a file, then takes it from there. The loop is tuned so a human-in-the-loop session still moves fast, often faster than reading Reddit yourself.

Do I need to know how to code?

No. It's a quick, guided workflow, not a coding task. The agent points you to exactly what to save and does the analysis for you. The included scripts just normalize the threads you saved, and the agent runs them for you.

Will it make up quotes?

No, and it's enforced, not just promised. When the report is drafted, a separate verifier agent checks every quote against the threads you saved and confirms the exact text is there. Anything it can't match to a real, linkable source is fixed, dropped, or flagged. That verification pass is a core part of the pack.

Which AI tool do I need?

Any AI coding tool that can read and write files on your machine: Claude Code, Codex, Cursor, and similar. It does not run in a plain web chat (Claude.ai or ChatGPT in the browser), because the agent has to save and read thread files locally. You point your tool at the agent and go.

Can I use it for client work?

Yes. The license covers research you personally perform, including for clients. You just can't resell or redistribute the pack itself.

Refunds?

Because this is an instant digital download, all sales are final. If something isn't working for you, reach out and I'll do my best to help.

The trust layer

Every quote is checked before you see it

When the report is drafted, the pack automatically spawns a separate verifier agent with a clean context. Its one job is to confirm that every quote in the report appears in the thread files you saved. Anything it can't find in the source is fixed, dropped, or flagged. You get research you can stand behind, with citations that hold up when someone clicks them.

Find out what your customers actually want

The Reddit Research Agent, from the creator of the Ivy League course, Startup Studio. Instant download, yours forever, $10.

Get my research agent