How long does a custom dataset take to deliver?

Most custom datasets are collected, enriched, and delivered within 1–3 business days. You will receive an email notification when your dataset is ready.

How are platforms selected for my custom dataset?

You pick a category and we auto-select the most relevant platforms for it from 100+ supported sources (Reddit, YouTube, GitHub, Mastodon, Bluesky, Telegram, Medium, and more). You can review and trim up to 20% before submitting.

What formats are custom datasets delivered in?

Custom datasets are delivered in CSV, JSON, JSONL, and Parquet formats. All formats are included at no extra cost.

How much do custom datasets cost?

Custom datasets are priced by row count — from $18 (up to 25K rows) to $68 (up to 100K rows). You'll see one clear total before you pay. Pro subscribers get custom datasets included at no extra charge.

What enrichment fields are included?

Rows are enriched with the fields relevant to your platforms, drawn from 250+ available across 30 categories: sentiment analysis, emotion detection, topic classification, entity extraction, geo/location, request intent, financial signals, engagement metrics, quality scores, and more.

Can I request private or exclusive datasets?

Yes. You can request private datasets that are not shared with other users. Private datasets include a privacy fee of $11.98 on top of the row-based pricing.

Custom Datasets — Social Intel

Request a custom social media dataset from 115 platforms. Pick platforms, keywords, and enrichment fields from 250+ available. Delivered in 1-3 days as CSV, JSON, JSONL, or Parquet. From $18.

Supported Platforms:
- GitHub (github)
- Hacker News (hackernews)
- Stack Overflow (stackoverflow)
- GitLab (gitlab)
- Bitbucket (bitbucket)
- DEV (devto)
- SourceForge (sourceforge)
- Hashnode (hashnode)
- Lobsters (lobsters)
- Launchpad (launchpad)
- npm (npm)
- Docker Hub (dockerhub)
- crates.io (cratesio)
- NuGet (nuget)
- Packagist (packagist)
- Pub.dev (pubdev)
- RubyGems (rubygems)
- Mastodon (mastodon)
- Lemmy (lemmy)
- Bluesky (bluesky)
- Misskey (misskey)
- PeerTube (peertube)
- Pixelfed (pixelfed)
- Pleroma (pleroma)
- Odysee (odysee)
- YouTube (youtube)
- Twitch (twitch)
- BitChute (bitchute)
- SoundCloud (soundcloud)
- Bilibili (bilibili)
- Discogs (discogs)
- Invidious (invidious)
- Last.fm (lastfm)
- Steam Community (steamcommunity)
- TMDB (tmdb)
- Coub (coub)
- Dailymotion (dailymotion)
- Funkwhale (funkwhale)
- Mixcloud (mixcloud)
- Newgrounds (newgrounds)
- WhoSampled (whosampled)
- MusicBrainz (musicbrainz)
- Reddit (reddit)
- Tumblr (tumblr)
- 4chan (4chan)
- 8kun (8kun)
- Minds (minds)
- Telegram (telegram)
- Diaspora (diaspora)
- Friendica (friendica)
- WriteFreely (writefreely)
- BookWyrm (bookwyrm)
- Micro.blog (microblog)
- Snapchat (snapchat)
- Dribbble (dribbble)
- Behance (behance)
- DeviantArt (deviantart)
- Imgur (imgur)
- Pinterest (pinterest)
- ArtStation (artstation)
- Pixiv (pixiv)
- Are.na (arena)
- 500px (500px)
- Roblox (roblox)
- CoinMarketCap (coinmarketcap)
- CoinGecko (coingecko)
- CoinDesk (coindesk)
- StockTwits (stocktwits)
- DexScreener (dexscreener)
- Pump.fun (pump.fun)
- Steam (steam)
- Trustpilot (trustpilot)
- Seeking Alpha (seekingalpha)
- G2 (g2)
- Capterra (capterra)
- Kickstarter (kickstarter)
- Indie Hackers (indiehackers)
- Product Hunt (producthunt)
- Goodreads (goodreads)
- Trakt (trakt)
- Letterboxd (letterboxd)
- LibraryThing (librarything)
- Medium (medium)
- Steemit (steemit)
- Substack (substack)
- The Guardian (theguardian)
- WordPress (wordpress)
- MetaFilter (metafilter)
- Archive of Our Own (archiveofourown)
- LiveJournal (livejournal)
- Locals (locals)
- Something Awful (something_awful)
- FanFiction.net (fanfiction)
- Flipboard (flipboard)
- Flickr (flickr)
- Unsplash (unsplash)
- arXiv (arxiv)
- Wikipedia (wikipedia)
- Wikidata (wikidata)
- BoardGameGeek (boardgamegeek)
- Chess.com (chesscom)
- Epic Games (epicgames)
- Jikan (jikan)
- Speedrun.com (speedruncom)
- HowLongToBeat (howlongtobeat)
- RAWG (rawg)
- OpenCritic (opencritic)
- Giant Bomb (giantbomb)
- MangaDex (mangadex)
- AniList (anilist)
- Kitsu (kitsu)
- Anime News Network (animenewsnetwork)
- Wattpad (wattpad)
- Anime-Planet (animeplanet)
- Royal Road (royalroad)


Enrichment Field Groups:

Core (27 fields):
- id
- content
- url
- title
- author
- author_id
- author_followers
- author_verified
- collected_at
- published_at
- likes
- comments
- shares
- views
- engagement_rate
- word_count
- reading_time_minutes
- duration_seconds
- content_snippet
- description_snippet
- region
- language
- posting_hour
- posting_day
- posting_period
- is_weekend
- is_business_hours

Sentiment & Quality (23 fields):
- sentiment_label
- sentiment_score
- sentiment_positive
- sentiment_negative
- sentiment_neutral
- sentiment_confidence
- sentiment_method
- sarcasm_flag
- sarcasm_score
- irony_type
- emotion_primary
- emotion_secondary
- emotion_scores
- language_detected
- text_quality_score
- toxicity_score
- formality_score
- subjectivity_score
- intensity_score
- argument_quality
- evidence_quality
- source_credibility
- tone

Engagement (10 fields):
- stars
- forks
- watchers
- claps
- responses
- open_issues
- view_velocity
- growth_velocity
- cross_region_appeal
- virality_score

Topics (5 fields):
- topic_tags
- enriched_at
- community_health_score
- original_platform
- duplicate_hash

Financial (17 fields):
- tickers
- ticker_count
- financial_signal
- financial_score
- has_price_target
- volatility_signal
- is_fud
- is_fomo
- has_dd
- finance_topic
- asset_class
- investment_strategy
- target_demographic
- goal_type
- debt_topic
- fire_mentioned
- inflation_signal

Jobs & Career (12 fields):
- salary_mentioned
- salary_min
- salary_max
- yoe
- seniority
- remote_policy
- role_type
- tech_stack
- company_size
- job_title
- company_mentioned
- visa_sponsorship_mentioned

Content Classification (17 fields):
- is_opinion
- has_sources
- reading_level
- is_practical
- writing_style
- content_format
- actionability_score
- has_citations
- has_code
- skill_level
- has_data
- audience_level
- is_case_study
- has_docs
- has_tests
- is_production_ready
- maintenance_status

Technical (9 fields):
- language
- license
- category
- tags
- topics
- is_fork
- open_issues
- repo_name
- primary_language

Video (7 fields):
- title
- duration_seconds
- description_snippet
- view_velocity
- region
- cross_region_appeal
- closed_captions_available

Article (8 fields):
- title
- reading_time_minutes
- word_count
- claps
- responses
- content_snippet
- has_citations
- has_data

Crypto & DeFi (10 fields):
- coins_mentioned
- coin_count
- pump_signal
- exchange_mentioned
- defi_protocol_mentioned
- nft_mentioned
- whale_activity_signal
- on_chain_events
- on_chain_primary_event
- on_chain_signal_strength

Politics & Civic (9 fields):
- political_lean
- policy_topic
- stance
- party_mentioned
- candidate_mentioned
- argument_quality
- misinformation_flag
- mobilization_signal
- civic_topic

Gaming (9 fields):
- game_title
- game_genre
- platform_type
- is_review
- hype_score
- content_type
- esports_event
- patch_mentioned
- release_signal

AI & ML (9 fields):
- models_mentioned
- techniques_mentioned
- frameworks_mentioned
- is_paper_discussion
- is_benchmark
- technical_depth
- ml_topic
- open_source_flag
- alignment_topic

Health & Wellness (9 fields):
- condition_type
- wellness_topic
- support_type
- is_professional_content
- health_has_scientific_ref
- evidence_quality
- mental_health_flag
- medication_mentioned
- is_progress_post

Climate & Environment (8 fields):
- climate_topic
- solution_type
- climate_has_scientific_ref
- policy_mentioned
- carbon_signal
- renewable_energy_type
- esg_signal
- corporate_target_mentioned

Creator Economy (7 fields):
- creator_platform
- monetization_type
- follower_tier
- revenue_model
- sponsorship_mentioned
- niche_topic
- collab_mentioned

E-commerce & Reviews (16 fields):
- product_mentioned
- price_tier
- business_model
- business_stage
- review_rating
- verified_purchase
- pros_mentioned
- cons_mentioned
- management_response
- niche_category
- refund_signal
- review_helpful_count
- review_response
- review_recommend
- review_title
- rating_boost

Fediverse (10 fields):
- instance_domain
- federation_protocol
- software_type
- boost_count
- reply_count
- visibility_scope
- content_warning
- hashtags
- mentioned_users
- local_only_flag

Live Streaming (6 fields):
- stream_platform
- viewer_peak
- chat_velocity
- subscriber_count
- stream_category
- vod_available

Meme & Viral Culture (5 fields):
- meme_format
- origin_platform
- spread_velocity
- cultural_reference
- reaction_type

Misinformation & Narratives (9 fields):
- claim_type
- verified_flag
- debunked_flag
- source_credibility
- echo_chamber_signal
- bot_signal
- coordinated_inauthentic_flag
- narrative_type
- fact_check_url

Extremism & Safety (6 fields):
- ideology_type
- target_group
- hate_signal
- recruitment_signal
- propaganda_format
- platform_ban_signal

Quality & Safety (8 fields):
- is_phishing
- is_spam
- low_relevance
- is_nsfw
- quality_score
- contradictions
- content_age_days
- is_stale_content

Composite Scores (7 fields):
- overall_score
- topic_coherence
- professional_score
- thought_leader_score
- tutorial_quality_score
- dev_hype_score
- pro_recovery_flag

Entity Extraction (5 fields):
- entities_people
- entities_organizations
- entities_locations
- entities_products
- entity_count

Bot Detection & Authenticity (4 fields):
- bot_likelihood_score
- bot_signals
- engagement_authenticity_score
- authenticity_signals

Geo & Location (6 fields):
- locations_mentioned
- countries_mentioned
- cities_mentioned
- regions_mentioned
- primary_geo
- geo_granularity

Question & Request Intent (4 fields):
- is_question
- question_type
- request_type
- question_markers

Statistics & Data Density (4 fields):
- has_statistics
- statistics_type
- numeric_density
- statistics_keywords


Pricing Tiers:
- 10K–25K rows: $18
- 25K–50K rows: $28
- 50K–75K rows: $48
- 75K–100K rows: $68


Frequently Asked Questions:
- Q: How long does it take to get a custom dataset?
  A: Typically 1-3 business days from submission to delivery. Complex requests with many platforms or high row limits may take slightly longer.
- Q: What formats are available?
  A: CSV, JSON, JSONL, and Parquet — all with standardized schemas and consistent column ordering.
- Q: Can I get data from a platform not listed?
  A: Yes. Contact us with your specific requirements. We support 100+ platforms and can often source data from unlisted platforms on request.
- Q: How many enrichment fields can I include?
  A: Choose from 250+ enrichment fields across 30+ categories. Core fields are always included; you can optionally add sentiment, emotion, financial signals, entity extraction, and more.
- Q: Is the data deduplicated?
  A: Yes — we apply SHA-256 content hashing and 5 dedup strategies at export.
- Q: What is the minimum row count?
  A: 10,000 rows minimum. Pricing scales linearly with row count in 25K-row tiers from $18 to $68.

Dataset Request

Request a new dataset

Tell us what you need. We'll collect, enrich, and deliver your private dataset — with a 100-row free sample and permanent archive access.

Request for review

We'll collect your dataset and notify you via email. The full download requires an individual purchase (from $18), with permanent access. 100-row free sample always included.

Free

Build progress20%

Complete the form to build your dataset — 80% to go.

Dataset Identity

Dataset Name *— used to generate slug

Platforms to Collect

Auto-selected from category— 0 selected · remove up to 1

Platforms are chosen automatically from your category. You can remove up to 20% (1) of them — collection always stays on-topic.

Fields to Include

Choose fields *— 15 selected

Methodology Documentation

15 fields selected

idcontenturltitleauthorauthor_idauthor_followersauthor_verifiedcollected_atpublished_atlikescomments+3 more

Row Count

Choose your row tier— determines price

10K–25K rows$18

10K25K50K100K

100-row free sample included. Pay once for full download — permanent access.

Privacy

Add this dataset to the public catalog

No privacy fee. Dataset is added to the public catalog for others to discover.

By submitting you agree to our Terms and Privacy Policy

Submitted requests appear in your Dashboard under Custom Datasets

Your dataset

NameUntitled dataset

Categorygeneral

Fields15 configured

Rows10K–25K rows · $18

FormatsCSV · JSON · JSONL · Parquet

Build progress20%

80% left — finish your spec.

Pricing

Selected tier: $18 — pay once, permanent access. Pro plan — 7-day free trial — includes free custom datasets.

Enrichment Pipeline

Every custom dataset goes through our deterministic enrichment pipeline, which draws on 250+ available fields and applies the ones relevant to its platforms.

1. Auto-Select Platforms

Platforms are auto-selected from your chosen category across 100+ supported sources — you can review and trim up to 20% before submitting.

2. Generate Dataset Description

AI writes a concise 1-2 sentence description of the dataset based on your input, category, and selected fields. Used for catalog display and metadata.

3. Expand Search Queries

AI generates optimized search queries that scale with your selected row tier — from ~65 at the 10K–25K tier up to 400+ at 75K–100K — covering multiple angles: tutorials, comparisons, opinions, news, how-tos, best-of lists, and trends. For large counts, queries are generated in parallel batches to guarantee full coverage without truncation.

4. Optimize Field Selection

15 CORE fields (id, content, url, author, engagement metrics, etc.) are always included for data quality and consistency. Additional field groups auto-suggest based on description keywords — sentiment, financial, technical, health, gaming, and more.

Enrichment Pipeline

Smart Scheduling: Collection tasks are prioritised and queued. Each platform is scraped using custom-built adapters.

Post-Level Enrichment: Each collected post is analysed for sentiment, emotion, topics, financial signals, and domain-specific features (crypto tickers, game titles, health conditions, etc.).

Deduplication: Content is deduplicated by hash across platforms to ensure clean, unique data.

Multi-Format Export: Delivered as CSV, JSON, JSONL, and Parquet with schema documentation.

Deterministic enrichment — every field reproducible from source code.

SocialINTEL

Fetching datasets…

Custom Datasets — Social Intel Request a custom social media dataset from 115 platforms. Pick platforms, keywords, and enrichment fields from 250+ available. Delivered in 1-3 days as CSV, JSON, JSONL, or Parquet. From $18. Supported Platforms: - GitHub (github) - Hacker News (hackernews) - Stack Overflow (stackoverflow) - GitLab (gitlab) - Bitbucket (bitbucket) - DEV (devto) - SourceForge (sourceforge) - Hashnode (hashnode) - Lobsters (lobsters) - Launchpad (launchpad) - npm (npm) - Docker Hub (dockerhub) - crates.io (cratesio) - NuGet (nuget) - Packagist (packagist) - Pub.dev (pubdev) - RubyGems (rubygems) - Mastodon (mastodon) - Lemmy (lemmy) - Bluesky (bluesky) - Misskey (misskey) - PeerTube (peertube) - Pixelfed (pixelfed) - Pleroma (pleroma) - Odysee (odysee) - YouTube (youtube) - Twitch (twitch) - BitChute (bitchute) - SoundCloud (soundcloud) - Bilibili (bilibili) - Discogs (discogs) - Invidious (invidious) - Last.fm (lastfm) - Steam Community (steamcommunity) - TMDB (tmdb) - Coub (coub) - Dailymotion (dailymotion) - Funkwhale (funkwhale) - Mixcloud (mixcloud) - Newgrounds (newgrounds) - WhoSampled (whosampled) - MusicBrainz (musicbrainz) - Reddit (reddit) - Tumblr (tumblr) - 4chan (4chan) - 8kun (8kun) - Minds (minds) - Telegram (telegram) - Diaspora (diaspora) - Friendica (friendica) - WriteFreely (writefreely) - BookWyrm (bookwyrm) - Micro.blog (microblog) - Snapchat (snapchat) - Dribbble (dribbble) - Behance (behance) - DeviantArt (deviantart) - Imgur (imgur) - Pinterest (pinterest) - ArtStation (artstation) - Pixiv (pixiv) - Are.na (arena) - 500px (500px) - Roblox (roblox) - CoinMarketCap (coinmarketcap) - CoinGecko (coingecko) - CoinDesk (coindesk) - StockTwits (stocktwits) - DexScreener (dexscreener) - Pump.fun (pump.fun) - Steam (steam) - Trustpilot (trustpilot) - Seeking Alpha (seekingalpha) - G2 (g2) - Capterra (capterra) - Kickstarter (kickstarter) - Indie Hackers (indiehackers) - Product Hunt (producthunt) - Goodreads (goodreads) - Trakt (trakt) - Letterboxd (letterboxd) - LibraryThing (librarything) - Medium (medium) - Steemit (steemit) - Substack (substack) - The Guardian (theguardian) - WordPress (wordpress) - MetaFilter (metafilter) - Archive of Our Own (archiveofourown) - LiveJournal (livejournal) - Locals (locals) - Something Awful (something_awful) - FanFiction.net (fanfiction) - Flipboard (flipboard) - Flickr (flickr) - Unsplash (unsplash) - arXiv (arxiv) - Wikipedia (wikipedia) - Wikidata (wikidata) - BoardGameGeek (boardgamegeek) - Chess.com (chesscom) - Epic Games (epicgames) - Jikan (jikan) - Speedrun.com (speedruncom) - HowLongToBeat (howlongtobeat) - RAWG (rawg) - OpenCritic (opencritic) - Giant Bomb (giantbomb) - MangaDex (mangadex) - AniList (anilist) - Kitsu (kitsu) - Anime News Network (animenewsnetwork) - Wattpad (wattpad) - Anime-Planet (animeplanet) - Royal Road (royalroad) Enrichment Field Groups: Core (27 fields): - id - content - url - title - author - author_id - author_followers - author_verified - collected_at - published_at - likes - comments - shares - views - engagement_rate - word_count - reading_time_minutes - duration_seconds - content_snippet - description_snippet - region - language - posting_hour - posting_day - posting_period - is_weekend - is_business_hours Sentiment & Quality (23 fields): - sentiment_label - sentiment_score - sentiment_positive - sentiment_negative - sentiment_neutral - sentiment_confidence - sentiment_method - sarcasm_flag - sarcasm_score - irony_type - emotion_primary - emotion_secondary - emotion_scores - language_detected - text_quality_score - toxicity_score - formality_score - subjectivity_score - intensity_score - argument_quality - evidence_quality - source_credibility - tone Engagement (10 fields): - stars - forks - watchers - claps - responses - open_issues - view_velocity - growth_velocity - cross_region_appeal - virality_score Topics (5 fields): - topic_tags - enriched_at - community_health_score - original_platform - duplicate_hash Financial (17 fields): - tickers - ticker_count - financial_signal - financial_score - has_price_target - volatility_signal - is_fud - is_fomo - has_dd - finance_topic - asset_class - investment_strategy - target_demographic - goal_type - debt_topic - fire_mentioned - inflation_signal Jobs & Career (12 fields): - salary_mentioned - salary_min - salary_max - yoe - seniority - remote_policy - role_type - tech_stack - company_size - job_title - company_mentioned - visa_sponsorship_mentioned Content Classification (17 fields): - is_opinion - has_sources - reading_level - is_practical - writing_style - content_format - actionability_score - has_citations - has_code - skill_level - has_data - audience_level - is_case_study - has_docs - has_tests - is_production_ready - maintenance_status Technical (9 fields): - language - license - category - tags - topics - is_fork - open_issues - repo_name - primary_language Video (7 fields): - title - duration_seconds - description_snippet - view_velocity - region - cross_region_appeal - closed_captions_available Article (8 fields): - title - reading_time_minutes - word_count - claps - responses - content_snippet - has_citations - has_data Crypto & DeFi (10 fields): - coins_mentioned - coin_count - pump_signal - exchange_mentioned - defi_protocol_mentioned - nft_mentioned - whale_activity_signal - on_chain_events - on_chain_primary_event - on_chain_signal_strength Politics & Civic (9 fields): - political_lean - policy_topic - stance - party_mentioned - candidate_mentioned - argument_quality - misinformation_flag - mobilization_signal - civic_topic Gaming (9 fields): - game_title - game_genre - platform_type - is_review - hype_score - content_type - esports_event - patch_mentioned - release_signal AI & ML (9 fields): - models_mentioned - techniques_mentioned - frameworks_mentioned - is_paper_discussion - is_benchmark - technical_depth - ml_topic - open_source_flag - alignment_topic Health & Wellness (9 fields): - condition_type - wellness_topic - support_type - is_professional_content - health_has_scientific_ref - evidence_quality - mental_health_flag - medication_mentioned - is_progress_post Climate & Environment (8 fields): - climate_topic - solution_type - climate_has_scientific_ref - policy_mentioned - carbon_signal - renewable_energy_type - esg_signal - corporate_target_mentioned Creator Economy (7 fields): - creator_platform - monetization_type - follower_tier - revenue_model - sponsorship_mentioned - niche_topic - collab_mentioned E-commerce & Reviews (16 fields): - product_mentioned - price_tier - business_model - business_stage - review_rating - verified_purchase - pros_mentioned - cons_mentioned - management_response - niche_category - refund_signal - review_helpful_count - review_response - review_recommend - review_title - rating_boost Fediverse (10 fields): - instance_domain - federation_protocol - software_type - boost_count - reply_count - visibility_scope - content_warning - hashtags - mentioned_users - local_only_flag Live Streaming (6 fields): - stream_platform - viewer_peak - chat_velocity - subscriber_count - stream_category - vod_available Meme & Viral Culture (5 fields): - meme_format - origin_platform - spread_velocity - cultural_reference - reaction_type Misinformation & Narratives (9 fields): - claim_type - verified_flag - debunked_flag - source_credibility - echo_chamber_signal - bot_signal - coordinated_inauthentic_flag - narrative_type - fact_check_url Extremism & Safety (6 fields): - ideology_type - target_group - hate_signal - recruitment_signal - propaganda_format - platform_ban_signal Quality & Safety (8 fields): - is_phishing - is_spam - low_relevance - is_nsfw - quality_score - contradictions - content_age_days - is_stale_content Composite Scores (7 fields): - overall_score - topic_coherence - professional_score - thought_leader_score - tutorial_quality_score - dev_hype_score - pro_recovery_flag Entity Extraction (5 fields): - entities_people - entities_organizations - entities_locations - entities_products - entity_count Bot Detection & Authenticity (4 fields): - bot_likelihood_score - bot_signals - engagement_authenticity_score - authenticity_signals Geo & Location (6 fields): - locations_mentioned - countries_mentioned - cities_mentioned - regions_mentioned - primary_geo - geo_granularity Question & Request Intent (4 fields): - is_question - question_type - request_type - question_markers Statistics & Data Density (4 fields): - has_statistics - statistics_type - numeric_density - statistics_keywords Pricing Tiers: - 10K–25K rows: $18 - 25K–50K rows: $28 - 50K–75K rows: $48 - 75K–100K rows: $68 Frequently Asked Questions: - Q: How long does it take to get a custom dataset? A: Typically 1-3 business days from submission to delivery. Complex requests with many platforms or high row limits may take slightly longer. - Q: What formats are available? A: CSV, JSON, JSONL, and Parquet — all with standardized schemas and consistent column ordering. - Q: Can I get data from a platform not listed? A: Yes. Contact us with your specific requirements. We support 100+ platforms and can often source data from unlisted platforms on request. - Q: How many enrichment fields can I include? A: Choose from 250+ enrichment fields across 30+ categories. Core fields are always included; you can optionally add sentiment, emotion, financial signals, entity extraction, and more. - Q: Is the data deduplicated? A: Yes — we apply SHA-256 content hashing and 5 dedup strategies at export. - Q: What is the minimum row count? A: 10,000 rows minimum. Pricing scales linearly with row count in 25K-row tiers from $18 to $68.

Enrichment Pipeline

Smart Scheduling: Collection tasks are prioritised and queued. Each platform is scraped using custom-built adapters.

Post-Level Enrichment: Each collected post is analysed for sentiment, emotion, topics, financial signals, and domain-specific features (crypto tickers, game titles, health conditions, etc.).

Deduplication: Content is deduplicated by hash across platforms to ensure clean, unique data.

Multi-Format Export: Delivered as CSV, JSON, JSONL, and Parquet with schema documentation.