What is a Listener?

A Listener is a saved set of instructions that tells Earshot which Reddit communities to watch and what to look for. Think of it as a pair of ears pointed at the conversations you care about.

Each Listener:

Monitors specific subreddits (Reddit communities)
Filters posts and comments using keywords and other rules
Runs on a schedule you set (every N minutes)
Surfaces matches as drafts (Solo) or monitored posts (Scout)

You can have multiple Listeners — for example, one watching founder communities and another watching developer tool discussions.

The goal isn't maximum drafts. It's high-intent threads where you can genuinely help. Tuning is iterative — expect to adjust weekly for the first few weeks.

Screenshot: Listeners list page showing a listener card with status and health metrics — ScreenshotListeners list page showing a listener card with status and health metrics

Creating a Listener

Go to Listeners → New listener. The form walks you through five sections.

Listener basics

Field	What it does
Listener name	A label to tell this listener apart from others you create.
Broad topics to focus on	Topics and angles the AI should care about when picking posts to reply to.

Screenshot: New listener form showing Listener name and Broad topics to focus on fields — ScreenshotNew listener form showing Listener name and Broad topics to focus on fields

Click Next. Earshot suggests communities based on your topic text.

Communities

Suggested communities — click Suggest subreddits to refresh ideas from your focus text, then Add the ones that fit.
Search subreddits — type to search the catalog (communities with 10k+ members). Pick up to your plan's subreddit limit per Listener.
Community rules — for each selected subreddit, expand Community rules — r/{name} to read moderator rules. This will give you an idea on what's allowed in each subreddit BEFORE you start listening.

Keyword configuration

Field	What it does
Primary keywords	A post or comment must mention at least one of these terms to be considered. Required — at least one.
Posts must also include	Optional extra terms. Choose Any of (match any secondary term) or All of (must match every secondary term).
Negative terms	Exclude posts that mention both a primary keyword and any of these terms.
Negative authors	Skip posts from these Reddit usernames for this listener only. Enter without `u/`.
Exact match	Match standalone terms, domains, and handles — not embedded substrings. On by default.
Case sensitive	Useful for acronyms like RAG that should not match "rag". Off by default.

These rules apply only to this Listener. For account-wide author exclusions, see Global keywords below.

Tips & tricks

Primary keywords — start with terms that signal real intent (put around 15-25):

Problem-space: "CI/CD", "monitoring tool"
Recommendation seekers: "looking for", "recommend", "best tool for"
Evaluation: "vs", "compared to", "switching from"

Posts must also include — pair a broad primary with intent terms. Example: primary "project management" + Any of "recommend", "alternative", "looking for" = recommendation threads only.

Negative terms — filter noise when a primary keyword still matches junk:

Primary: your product category. Negative: "free", "open source only", "hobby project" (if those aren't your audience).
Primary: competitor name. Negative: "meme", "joke", "shitpost".

Exact match (on by default) — "RAG" won't match "drag". Turn on Case sensitive for acronyms that shouldn't match lowercase words.

Screenshot: Listener keyword configuration with Primary keywords and Negative terms — ScreenshotListener keyword configuration with Primary keywords and Negative terms

Global keywords

Settings → Global keywords blocks authors account-wide:

Level	Where	Scope
Per-listener	Negative authors on the Listener form	This Listener only
Account-wide	Settings → Global keywords → Negative authors	All Listeners

Help text on the settings page: "Posts from these authors are skipped across all listeners." There are no global keyword inclusions — only author exclusions at the account level.

Posts and comments

In the app this section is labeled: What to watch for.

Field	What it does
Watch for	Comments, Posts, or Both.
Check every (minutes)	How often this Listener runs. Minimum 10 minutes; maximum 24 hours (1440 minutes). Click Suggest for a recommendation based on subreddit activity.
How deep in the thread?	Shown when watching comments. Top-level comments only or I'm okay with more depth (replies to top-level comments).

On Solo, you'll also see Writing voice (attach a Persona) and Knowledge bases (attach reference docs) before the final section. Both are optional. Scout hides these cards.

Fine tuning

In the app this section is a collapsible card titled Fine-tuning.

Field	What it does
How strict should filtering be?	Loose lets more through; Medium is balanced (default); Tight is more selective.
Minimum comment length	Ignore comments shorter than this many words. Default: 8. (the default skips short comments which don't add much value)
Skip downvoted comments	Don't draft replies for comments with a negative score. On by default. (the default skips downvoted comments which are likely not worth replying to)

Click Create listener. You'll land on the edit page for your new Listener.

Editing a Listener

On the edit page, all sections are visible at once. The main difference from create:

Active toggle — Run this listener on schedule. Turn off to pause scheduled checks.
Attached knowledge bases — Remove existing links or Attach new ones with the Add knowledge base dropdown.

Save your changes when done.

Listener runs

Listeners run automatically on the schedule you set.

On each listener card you'll see:

Next check in {time} — when the next scheduled run is due
Last checked {time} ago — when the last run finished
Paused — if you've turned off the Active toggle
Checking soon — if the Listener has never run yet

While a run is in progress, you'll see Running pipeline… with a progress bar.

Health funnel

Each card shows a Health section with a funnel:

Scanned → Keyword filter → Contextual filter → Drafts (Solo) or Matches (Scout) → Posted

Pattern	Likely cause	Fix
High scanned, low keyword filter	Keywords too broad or too few	Add negative terms, tighten primaries
High keyword, low contextual	Strictness too tight or focus text too narrow	Loosen strictness, broaden your topic description
High contextual, low drafts	(shouldn't occur ideally) Solo draft limit or processing issue	Check daily limit
High drafts, low posted	You're rejecting a lot (that's OK!) or not posting yet	Review draft quality, tune Persona/KB

Screenshot: Listener card health funnel showing Scanned, Keyword filter, Contextual filter, and Drafts or Matches stages — ScreenshotListener card health funnel showing Scanned, Keyword filter, Contextual filter, and Drafts or Matches stages

Run history

Expand Run history on a listener card to see what each scheduled check did — duration, items scanned, keyword and contextual filter counts, and drafts produced. Use this to tune keywords, strictness, and subreddits.

Scout vs Solo differences

Feature	Scout	Solo
Output label on card	Matches	Drafts
Writing voice (Persona)	Hidden	Available
Knowledge bases	Hidden	Available
Listener limit	Up to 3	Up to 10
Subreddits per Listener	Up to 6	Up to 10

What's next

Planning your Listeners — split business and personal Listeners
Personas — shape how Solo drafts sound
Drafts — review and post your matches
How Earshot works — how Listeners fit with Personas, KBs, and Drafts