Bluesky Brand Mention Scraper - Social Listening

Find out what people are saying about your brand on Bluesky, and know instantly which mentions actually need a reply.

This Actor does not just dump posts. Every mention comes back enriched with sentiment, question and complaint detection, a priority flag, and an engagement score, so a human (or an automation) can triage in seconds instead of reading everything.

Coverage limitation: read before running

Public post search can return HTTP 403 from datacenter and proxy IPs. By default, allowFeedFallback: true scans recent feeds of up to five accounts matching your keyword and keeps only posts containing that keyword. This is not network-wide monitoring and can miss most third-party mentions. Set allowFeedFallback: false to reject that reduced coverage.

Every row includes source (searchPosts or authorFeed) and coverage. The OUTPUT key-value-store record reports search errors, fallback usage and whether the budget stopped the run. A successful run means delivery, not complete coverage.

Generic scraper vs this Actor

Generic Bluesky scraper gives you	This Actor gives you
A raw dump of posts	A triaged mention feed
No sentiment signal	sentiment: positive / negative / neutral (heuristic)
You read every post to find questions	isQuestion flag for people asking about you
Complaints buried in the noise	isComplaint flag for bugs, outages, billing, frustration
No sense of what matters	priority: high on questions and complaints
Raw like/repost counts, if any	engagement score: likes + reposts + replies + quotes
Hashtags and links left inside the text	hashtags and links extracted per post
The same post repeated across keywords	Deduplication across keywords
Pay for everything scraped	Pay only for mentions returned after your filters

Example output (illustrative, not a measured customer post)

{
  "keyword": "apify",
  "text": "Is Apify down? Getting errors on every run",
  "sentiment": "negative",
  "isQuestion": true,
  "isComplaint": true,
  "priority": "high",
  "engagement": 14,
  "likeCount": 9, "repostCount": 2, "replyCount": 3, "quoteCount": 0,
  "hashtags": [], "links": [],
  "authorHandle": "user.bsky.social",
  "createdAt": "2026-07-23T10:00:00.000Z",
  "url": "https://bsky.app/profile/user.bsky.social/post/abc"
}

At a glance, in table form:

keyword	text	sentiment	isQuestion	isComplaint	priority	engagement	authorHandle
apify	Is Apify down? Getting errors on every run	negative	true	true	high	14	user.bsky.social

Output fields

Field	Type	Description
keyword	string	The monitored keyword that matched this mention
text	string	Full text of the Bluesky post
sentiment	string	positive, negative or neutral (heuristic, keyword-based)
isQuestion	boolean	True when the post looks like a question about the keyword
isComplaint	boolean	True when the post looks like a complaint (bug, outage, billing, frustration)
priority	string	high for questions and complaints, otherwise normal
engagement	integer	Total engagement: likes + reposts + replies + quotes
likeCount	integer	Number of likes
repostCount	integer	Number of reposts
replyCount	integer	Number of replies
quoteCount	integer	Number of quote posts
hashtags	array	Hashtags extracted from the post text
links	array	Links extracted from the post text
authorHandle	string	Bluesky handle of the author
createdAt	string	ISO 8601 timestamp of the post
url	string	Direct link to the post on bsky.app

How to use it

1. Click Try for free and open the Actor.
2. Run it with empty input {} to see it work with zero configuration, or enter your own keywords.
3. Optionally filter: set onlyQuestionsAndComplaints or raise minEngagement to cut noise.
4. Export the results to CSV, JSON, Excel, or pull them through the Apify API or a webhook into your own stack.

Input

{
  "keywords": ["apify", "my-brand"],
  "sort": "latest",
  "minEngagement": 0,
  "onlyQuestionsAndComplaints": false,
  "maxResultsPerKeyword": 100
}

Field	Type	Default	Description
keywords	array	["apify"]	Brands / keywords to monitor (each searched separately)
sort	string	latest	latest or top
minEngagement	integer	0	Only return mentions with at least this total engagement
onlyQuestionsAndComplaints	boolean	false	Only return high-priority mentions (questions + complaints)
maxResultsPerKeyword	integer	100	Max mentions per keyword (1-1000)

Example: only questions and complaints

{
  "keywords": ["my-brand"],
  "sort": "latest",
  "onlyQuestionsAndComplaints": true,
  "maxResultsPerKeyword": 200
}

Example: high-engagement mentions only

{
  "keywords": ["my-brand", "competitor-brand"],
  "sort": "top",
  "minEngagement": 10,
  "onlyQuestionsAndComplaints": false,
  "maxResultsPerKeyword": 100
}

Use cases

• Brand and reputation monitoring - triage the matching posts returned; coverage is not exhaustive.
• Customer support triage - route priority: high mentions straight into your support queue.
• Competitor tracking - watch rival brand names and how sentiment around them moves.
• Campaign measurement - track keyword and hashtag engagement across a launch.
• Lead generation - catch people publicly asking "which tool does X?" while the question is still open.

Integrations

Connect the output to Zapier, Make, n8n, Slack, Google Sheets, Airbyte, LangChain, the Apify REST API, or any webhook endpoint. A common setup: schedule the Actor hourly with onlyQuestionsAndComplaints: true and push each new high-priority mention into a Slack channel, so your team gets a steady, low-noise feed of only the posts worth replying to.

Pricing

Pay Per Event:

• $0.004 per enriched mention returned
• $0.00005 per Actor start

You only pay for mentions actually returned after your filters are applied. Tightening minEngagement or turning on onlyQuestionsAndComplaints lowers both the noise and the bill.

FAQ

Do I need a Bluesky account? No. The Actor reads public Bluesky AT Protocol endpoints (api.bsky.app) and requires no login or credentials.

Is this legal? It collects only publicly available posts, without logging in and without accessing private data. As always, use the results in line with your own local regulations and Bluesky's terms.

How fresh is the data? Results come from public search when reachable, otherwise from the limited account-feed fallback if enabled. Fallback ignores search ranking and does not guarantee the latest network-wide matches. Inspect source and coverage before using the results.

Why is the sentiment sometimes wrong? Sentiment, question and complaint detection are heuristic and keyword-based, tuned for English. They are designed as a triage signal to help you find what to look at first, not as ground truth. Sarcasm, slang and non-English posts can be misclassified.

Can I monitor competitors? Yes. Put competitor names or product terms into keywords; each keyword is searched separately and the results are deduplicated.

Will it break? Public endpoints can block requests. Retries and an explicitly marked limited fallback help with availability, but do not guarantee completeness. Disable fallback if your workflow requires real keyword search.

Troubleshooting

Problem	Fix
Empty dataset	Broaden your keywords and lower minEngagement to 0
Too much noise	Set onlyQuestionsAndComplaints to true and/or raise minEngagement
Missing recent posts	Increase maxResultsPerKeyword and set sort to latest
Same post appears under two keywords	Mentions are deduplicated across keywords; if you still see near-duplicates, they are distinct posts with similar text

Notes

• Source: public Bluesky AT Protocol endpoints (api.bsky.app). No login required.
• Sentiment, question and complaint detection is heuristic (keyword-based), tuned for English; treat it as a triage signal, not ground truth.