FAQs

The Siftsy Score is a composite 1–10 rating combining three dimensions: sentiment, relevance, and consensus. Sentiment measures the tone of the comments in relation to the content. Relevance measures how closely the topic of the comments relate to the post. Consensus measures how aligned commenters are. Above 5 is solid for branded content; 7–8 is strong; 8+ is exceptional. Read it alongside the underlying themes — it is a signal, not a verdict. For the full breakdown of how each dimension is calculated, refer to the Siftsy Score Methodology.

Comments are not read in isolation. Siftsy uses the post's caption, hashtags, creator bio, visuals, and video transcript (where available) as context — so it knows "that's fire" on a cooking video is praise, and an ironic "so helpful 🙄" is not. You can manually correct edge cases. See Can I manually correct a classification I disagree with? below.

Relevance measures how closely comments relate to the post's actual content, which includes whether or not a posts A post can have very positive comments that are entirely off-topic — high sentiment plus low relevance means you earned attention, but not on your intended message. Relevance is what separates genuine brand resonance from incidental virality.

Instagram, TikTok, YouTube, LinkedIn, X, and Facebook. Reddit, Threads, and Twitch are not yet supported. We prioritize new platforms by customer demand — tell us what matters most to you. See the full status and access breakdown in Supported Social Platforms.

Social listening is a telescope, scanning the whole landscape for volume, reach, and keyword mentions. Siftsy is a microscope, it goes deep on individual posts and tells you what the comments actually mean. They are complementary: use listening to find which content matters, then Siftsy to understand why the audience reacted as they did. See Is this just sentiment analysis? and What does Siftsy do that my social listening tool's built-in comment analysis doesn't?

You select the posts you want to analyze. Siftsy does not auto-select posts for you.

Paste individual post URLs, upload a batch via CSV, or use the account link feature to pull posts from a specific account. Many teams pair Siftsy with a social listening platform: find posts by virality, trend, or interest in your listening tool, then bulk upload those URLs to Siftsy to understand why the audience reacted as they did. See How to bulk upload posts.

Boosted posts (ads built on an organic post with comments enabled) work fully. Dark posts (ads with no public post) are not supported — they require advertiser-level access. Typical rule of thumb: if comments are publicly visible, Siftsy can analyze it.

Note: For TikTok, posts that are age-gated AND paid promotions are not treated as "public" by TikTok, and are not accessible at the moment. Updated July 2026

Sampling activates once a comment section exceeds your plan's base comment count. Below that threshold, every available comment is analyzed in full. Above it, Siftsy analyzes the top engaged comments plus a random 10% sample of the remainder — capturing both the loudest voices and the broader conversation.

This approach ensures a statistically significant sample of the conversation reguardless of comment section size, and extends analysis to any publicly available post equally.

For more details, see Understanding Comment Sampling.

About 5 minutes on average, though it varies with how many posts you submit and how many comments each has. The brief wait is deliberate: accurate sentiment scoring requires reading each comment in context and running it through AI processing — that is what lets Siftsy catch sarcasm, slang, and topic-vs-message nuance that faster keyword tools miss.

Scores are calibrated against real comment sections, where negativity and off-topic noise are common. In practice: 5–6 is average for branded content, 7–8 signals genuine resonance, 8+ is exceptional, and below 4 means a real disconnect. Build your own benchmarks over time — your campaign average matters more than any single post.

Siftsy collects as much context on a post as is available — caption, hashtags, creator bio, images, video, and transcripts — and derives the key context that drives scoring. That is what lets it judge whether a comment is relevant. This is the same engine behind sarcasm handling in How do you handle sarcasm, slang, and emojis?

Yes — text replies are analyzed alongside top-level comments, including nested reply threads. Video replies (for example on TikTok) are not yet included in text analysis but are on the roadmap.

Message sentiment is how the commenter feels overall (their tone). Topic sentiment is how the comment relates to the post's topic. "Ticketmaster needs to be stopped and I'm glad this artist is speaking out" is positive in message sentiment (happy, supportive) but negative on the topic of Ticketmaster. That split is critical for crisis and controversy contexts.

Yes — each refresh (weekly on Growth, daily on Powerhouse and Enterprise, plus any on-demand re-sync) captures the comment section's current state. View a post's score history to see how sentiment evolved — useful for watching a viral moment or PR situation play out over days or weeks. For refresh mechanics, see How to refresh and update post data and How often does the data refresh? below.

Yes. Siftsy works on any public post — just paste the URL. No account access, login, or permission needed. That makes it useful for competitive benchmarking, creator vetting, and new-business intelligence. For pitching and competitive analysis, see Can Siftsy be used for new-business pitches and competitive analysis? For UGC, see Can I analyze UGC at scale?

Yes. It is especially valuable for B2B brands and professional-services clients where LinkedIn is the primary channel. Analyze posts, track sentiment on thought leadership, and benchmark against competitors.

No — public posts only, with comments enabled. Private accounts and follower-restricted content cannot be analyzed, since we work through public APIs and licensed sources. This is also part of how we stay compliant with platform terms.

Not currently supported. We prioritize platform expansion by customer demand — tell us what is critical and we will factor it in. Today's platforms are Instagram, TikTok, YouTube, LinkedIn, X, and Facebook (public Page posts).

No special access required — paste the public post URL and Siftsy retrieves and analyzes the public comments. No connected account, admin rights, or owner permission needed. Applies equally to your content, competitors', and influencers'. See Can you analyze competitor and influencer posts we don't own?

Yes, as long as comments are publicly enabled and the post is reachable via the platform's public API. Where a platform restricts API access due to age-gating, we will verify case by case.

Yes — as long as posts are public, analyze them regardless of who created them. For large programs (thousands of posts per month), enterprise plans with higher credit limits and API access are the right fit. Also useful for auditing creator content quality before you amplify or boost it. See Can you analyze competitor and influencer posts we don't own?

Siftsy analyzes the comments on the stream post, but does not have access to live chat data. The live chat that occurs during a YouTube broadcast is separate from the comments on the post and is not captured by Siftsy.

The platform's top-line engagement metrics are reported by the platform and reflect a count of overall engagement. That count often does not match the number of comments actually available for analysis — platforms typically preserve the comment count even after comments are no longer visible.

Siftsy analyzes comments that are publicly visible. Comments that have been deleted, removed, or are otherwise not viewable by the general public fall outside that scope, even when they're still counted in a platform's top-line post metrics. Comments may be unavailable for several reasons:

• Deleted by the commenter

• Removed by the posting account

• Posted by accounts that have since been made private or deactivated

• Image-based comments, which have limited support

• Certain reply threads, which may not be fully captured

The test we apply is visibility to a normal logged-in account with no relationship to the post. If a comment can't be seen that way, we treat it as removed and it isn't captured — including comments visible only to their author.

Some comments are "hidden" are filtered out of a platform's default view without being removed — auto-filters, "view more" toggles, and collapsed reply threads. These are still reachable by any ordinary viewer, so they're captured and included in analysis.

Point-in-time data

Comment data reflects the state of the comment section at the moment the data was synced. Comments added after that point aren't included unless the post is re-synced. For active posts, this is the most common reason a report's comment count differs from what you see on the platform today. See How to refresh and update post data.

Sampling

There are two separate reasons a comment count may not line up: platform-side availability, and Siftsy's sampling. Sampling only activates once a comment section exceeds your tier's base comment count. Below that threshold, every available comment is analyzed — so any discrepancy is platform-side.

When sampling does activate, Siftsy constructs the sample in two parts:

• Base comments — the top engaged comments are included first

• Remaining comments — a random 10% sample of everything else

Sample size scales with comment volume up to the tier's maximum, so larger comment sections yield larger samples.

This approach guarantees the highest-engagement comments — the ones driving the conversation — are always included, then samples the long tail to keep proportions representative. It's optimized for directionally accurate insights into themes, sentiment, and share of voice, not guaranteed coverage of every individual comment. See Understanding Comment Sampling.

Full Reads (in development)

Full Reads will capture and process every available comment for selected analyses, using extra credits on an as-needed basis. It's designed for use cases that require complete comment coverage and exhaustive review.

If that fits your team's needs, reach out and we'll follow up when it's available.

Exports

Exports are capped on a per-post basis and never exceed the comments included in analysis. See How to export comments. If a comment section was sampled, the export reflects that sample.

Siftsy is built to interpret comment sections, not to redistribute them. This limit also keeps usage consistent with platform terms around bulk storage of comment data.

Yes. Paste a spreadsheet of post URLs and Siftsy processes them as a batch — ideal for large creator rosters, multi-post campaigns, or migrating from another tool. See How to bulk upload posts.

Smart search uses semantic AI to find comments about a concept, not just an exact word. Search "shipping delays" and it surfaces "took forever to arrive" or "still waiting on my order" even when "shipping" never appears — far better for surfacing themes in informal, varied language.

Yes — creator vetting is one of our strongest use cases. Analyze a creator's last 10–20 posts to see audience sentiment, comment quality, recurring themes, and whether their community is genuinely engaged or passive — signal that follower count and engagement rate alone cannot give you.

Posts live in campaigns — folders by client, creator, or time period. Add posts individually or via bulk CSV. Campaign-level aggregate scores give a campaign-wide view alongside individual post breakdowns. See How to create campaigns and How to edit campaigns.

Yes — filter by username or handle in the comment section below at the bottom of the analysis panel. See How to filter comments and How to explore and search the comments.

No. Siftsy does not surveil or track individual users, and we don't identify, rank, or follow specific commenters across posts. We analyze comment sections in aggregate. Within an analysis, you can filter comments by username or keyword but Siftsy doesn't build a profile of frequent commenters or show where a given user has commented.

Yes — a top use case for agencies. Analyze a prospect's own content (what their audience reaction looks like today) and their competitors' (what they are missing) before the first meeting. Many agencies build the pitch slide on Siftsy data before they have even met the client. See Can you analyze competitor and influencer posts we don't own?

Yes — add multiple creators' posts to one campaign to compare Siftsy Scores, sentiment breakdowns, and audience themes. Rank your roster by comment quality, spot which partnerships drive real conversation, and back casting decisions with data. See Can we evaluate creators before signing them?

Yes — track high-priority posts with daily refresh plus on-demand re-sync (30-minute minimum), filter by sentiment to surface the harshest reactions fast, and use smart search to find comments about a specific claim or issue. Score-threshold alerts are in development. See How often does the data refresh?

Yes — the top 40 languages globally, including Spanish, French, Portuguese, Arabic, German, Japanese, and Korean. The AI handles colloquial usage and regional slang, not just formal language. Each comment is scored in its native language. For the full list, refer to the Siftsy Score Methodology. For mixed-language posts, see Does Siftsy support multilingual content in a single post?

Yes — it auto-detects and processes multiple languages within one post, scoring each comment in its native language. You can filter or segment by language, and enterprise accounts get regional breakdowns to compare how markets respond to the same content. See Language Support and Does Siftsy support non-English languages?

Yes — override any classification directly in the platform. This matters most for niche terminology or insider community language. Corrections reflect immediately in your scores and theme breakdowns. See How to start analyzing comments.

No — Siftsy is a comment intelligence platform. Sentiment is only one of three score components alongside relevance and consensus. More importantly, we surface the themes driving the score, the comments illustrating them, and how sentiment breaks down by topic — the "why" behind the number, not just the number.

No direct native integrations yet. The current workflow is a CSV bridge — export post links from your existing tool and bulk-upload to Siftsy. Enterprise API access supports more seamless workflows, and native integrations are on the roadmap.

Most listening tools treat comments as a keyword signal — find a term, assign a broad label. Siftsy analyzes each comment's meaning in the context of its post: what is said, why, and how it relates to the brand message. The result is a structured intelligence layer — themes, consensus, relevance, quote-level examples — not a single positive or negative tag. The difference between knowing comments exist and understanding what they mean. See How does this compare to Meltwater, Brandwatch, or Sprinklr? and Is this just sentiment analysis?

You can buy ad hoc credits at $250 per 100 posts. They never expire while your subscription is active. We send overage alerts so you are never caught off guard. For high-variance volume (crisis, PR, big events), we suggest keeping a small buffer or discussing a flexible enterprise plan. See How to purchase a plan and How to track MAPs usage.

Siftsy standard plans are month-to-month with no long-term commitment. Enterprise plans require an extended commitment, and annual plans are available at a 15% discount. See How to purchase a plan or contact us.

Ad hoc credits ($250 per 100 posts) never expire while your subscription is active, so keep a small reserve for unpredictable moments. Enterprise plans can add flexible caps or event-based packages. We would rather you have coverage than hit a wall mid-crisis. See What happens if I exceed my monthly post credit limit?

Monthly plan credits reset at the start of each billing cycle and do not roll over. Ad hoc credits never expire while your subscription is active. If your usage is seasonal, consider a custom enterprise plan or building an ad hoc reserve during slower months. See What happens if I exceed my monthly post credit limit? and How to track MAPs usage.

No. Zero-comment posts show up in your account flagged as insufficient data for scoring, but consume no credit. Every other post uses exactly one credit (monthly or ad hoc) regardless of how many comments it has — whether it is 50 comments or 500,000. The one exception: re-syncing a post within your plan's included refresh period (daily, weekly, or none) uses one additional credit. See How often does the data refresh?

Yes — typical evaluation periods start at 30 days with 50 post analyses. Full platform access, so you can test on your own content, competitors, or creators.

Growth refreshes weekly; Powerhouse and Enterprise refresh daily. You can also trigger an on-demand re-sync of any post at any time (30-minute minimum between syncs). Each re-sync captures the latest metrics and comments as a new version — it never deletes existing comments, so you can compare versions over time. Re-syncing a post within your plan's included refresh period uses one additional credit per post. For step-by-step instructions, see How to refresh and update post data. See also Can I track sentiment changes over time on the same post? and Can Siftsy help with crisis monitoring and response?

24 months on Growth, 72 months (6 years) on Powerhouse and Enterprise. We retrieve the comments currently public on older posts — counts may be lower than at original posting if comments or accounts were removed since.

Yes — enterprise accounts support multiple users, teams, and workspaces. Organize by brand, region, or client, and set view, edit, and manage permissions. Included seats vary by plan — see How to purchase a plan for details. See How do I add team members and manage access?

From the Settings panel: invite by email and assign roles (admin, editor, viewer). Enterprise plans can restrict access to specific campaigns or workspaces. Read-only viewer access is available for clients who need to see reports but not edit. Included seats vary by plan — see How to purchase a plan for details. See How to add teammates to your team.

CSV, Excel, PDF, image, and plain text — covering summary scores, full comment breakdowns, and theme analyses. CSV and Excel include verbatim comment text. Word export and shareable no-login links are in development. Enterprise accounts also get programmatic export via API. See How to export comments and How to export insights.

Shareable no-login links are in development. For now, export reports as PDF, image, CSV, or plain text for client decks. Enterprise accounts can also add clients as read-only users. See How to export insights.

Yes, CSV and Excel exports include verbatim comment text alongside sentiment, relevance, and themes. Use it in your own reporting, custom analyses, or proprietary models. Enterprise API enables automated extraction at scale. See How to export comments.

Yes. While Siftsy intends to serve as an analysis platform to extract insights from the comments and not as a bulk comment exporter, you can do this should you need. You can filter the comments from positive to negative sentiment in the Explore tab, then select and copy the filtered comments as structured text. See How to export comments by sentiment for the full workflow.

We access social post and comment data through a combination of official platform APIs and licensed third-party data partners.