Help
56 articles
A run that fails, or a report with fewer than 200 relevant posts, returns its report automatically; the results stay visible, marked low data.
Unused one-off reports (Starter and Pack) are refundable within 14 days of purchase: write to support@senti.now.
Unused one-time add-ons are refundable within 14 days of purchase. Once any unit of a purchase is used, that purchase is no longer eligible for a voluntary refund, subject to your statutory rights. Write to support@senti.now.
Plans cancel anytime from Manage billing on your account page and stay active until the end of the paid period.
Plan reports are added each month and expire at the end of that month; unused ones do not roll over. Starter and Pack reports expire 365 days after purchase.
- Starter: 1 report: 3 months, four platforms; Change explanations: what changed, why, what to consider; Explorer, drill-down and share link; Unused reports refundable for 14 days.
- Pack: 5 reports, used any time within 365 days; Change explanations: what changed, why, what to consider; Side by side comparison; Unused reports refundable for 14 days.
- Pro: 6 reports a month; 3 tracked brands; Brands and campaigns update weekly; 3 creator posts a month; Change explanations: what changed, why, what to consider; Campaigns and alerts; Campaign impact report; CSV and print export.
- Agency: 10 reports a month; 5 tracked brands; Brands and campaigns update daily; 5 creator posts a month; Change explanations: what changed, why, what to consider; Campaign impact report; Live pulse and campaign war room; API access; Team access and client logins.
Every paid account can share reports, compare brands side by side, download charts as PNG and analyze people and topics (Beta). Agency also adds webhooks and your logo on PDF reports. See pricing.
- 5 reports (Pro and Agency): $19 for 5 reports, once. Valid 365 days, used after the month's reports.
- Extra tracked brand, updated weekly (Pro): $5 a month, or $50 a year on annual plans. One more brand, updated weekly like your plan's own.
- Extra tracked brand, updated daily (Agency): $12 a month, or $120 a year on annual plans. One more brand, updated daily like your plan's own.
- 5 creator posts (Pro and Agency): $9 for 5 creator posts, once. Five more creator posts on your campaigns, used within 365 days; each is followed for up to 30 days while your plan is active.
- Live-mode day for one brand (Agency): $9 a day, once. Live mode for one brand for one day, kept until used, for 365 days.
- Team seat (Agency): $10 a month, or $100 a year on annual plans. One more teammate on the account.
- Client logins (Agency): Included. Up to 10 read-only logins, each seeing only the brands you assign.
Choose them on the one screen before payment when you subscribe, or buy them later from your account page; Usage there shows what each limit has used. Recurring add-ons renew with your plan, monthly or yearly. A higher quantity applies once paid, charged for the rest of the period; a lower one applies from your next renewal. Reports, creator posts and live-mode days are one-time purchases. Refunds follow How do I get a refund?
Plans include scheduled refreshes, live pulses and live mode for normal use. Fair use is there for the rare account whose automatic use goes far beyond what its plan is meant for. Usage on your account page shows how much of this billing period's monitoring allowance is used, as a percentage.
When your automatic use in a billing period gets close to that point, we email you once and your brands page says that scheduled refreshes may pause soon if use continues. If it does go beyond, scheduled refreshes of your tracked brands pause, your brands page says so, and we email you. They resume at the start of your next billing period (your brands page says so once), or sooner if you write to support@senti.now.
Each extra tracked brand you add raises that allowance by what one more brand at its update schedule needs. Refreshes you start yourself on a tracked brand count toward it too, and wait for your next billing period once it is used up. New reports you start yourself are never affected, and your brands and their history stay as they are. Pausing tracking of a brand you no longer follow closely keeps automatic use down.
A report is one brand analyzed over up to three months on X, Instagram, Reddit and TikTok: the public posts we find are labeled for sentiment, topic and intent (very large brands are sampled evenly, see below), with the posts behind every number, a posts explorer and a receipt.
Adding a brand uses one report. Refreshing a brand that is not tracked uses one. Relabeling after you change topics is free.
Start on Analyze your brand: type the brand's name, and under More options its keyword rules, exclusions and languages. Nothing is fetched or charged yet, and it is saved for 7 days.
Then sign in (with Google or an email link, on any device). We search for the name and show you how many posts we found and sample posts that match your rules, before anything is charged. You confirm which brand you mean and what is analyzed, and only then you choose how to pay if you have no report left. The report starts as soon as the payment is confirmed; if you leave checkout, your brand stays saved and you can pay later.
Before your first purchase there are 2 checks: the first, then one more after you change the keyword rules; checking again after that comes with a report. The brand's industry and topics are set when its report starts.
Each line in the keywords field is one rule. A post matches if any line matches.
ipsy wellness: posts with every word, in any order, or the hashtag #ipsywellness."ipsy wellness": the exact phrase only.#ipsywellness,@ipsy: that hashtag or handle.ipsy wellness -sale: as above, without posts that say "sale".- The exclusions field removes posts containing any of its words, for every line.
Matching ignores case and works on whole words. A brand name of two words is searched as that phrase plus its hashtag and handle, never the words apart.
Every post is checked against these rules before it is labeled: one where no keyword appears as whole words ("ipsy" does not match "ipsybag") is dropped. The receipt shows how many were dropped; open the count to see examples and the reason.
Add words to the exclusions field: posts containing any of them are removed, for every keyword line. An exact phrase in quotes, a hashtag or a handle narrows the matches. Posts that turn out not to be about the brand are excluded from every number.
Fewer than 200 relevant posts were found for the brand. The results stay visible, marked low data, and the report is returned to your balance automatically.
When a platform returns a full first page, we collect one page per week and label an even sample of posts across each week, never only the newest. The receipt says so, for example "labeled 2,600 posts sampled evenly from about 345,000 found". See the methodology.
Each platform has its own typical sentiment, so the score can move when the share of posts from each platform changes, even if sentiment on every platform holds. For a change our significance test confirms, we split it into the change within platforms and the shift in which platforms the analyzed posts came from, and say which explains most of it, or that they cannot be told apart. Shares are of the posts we analyzed. See the methodology.
X is the freshest, and the live pulse on Agency reads X posts from the last minutes. Reddit and TikTok data can lag by hours; Instagram is collected one page per search. A platform that does not answer is marked unavailable on the report, the report is returned to your balance, and the platform can be fetched later at no report cost.
Reports read public posts on X, Reddit, TikTok and Instagram only. They do not cover news sites, blogs, forums, review sites, Facebook, YouTube, LinkedIn, podcasts, TV or print, which many listening tools include. Instagram is collected one page per search, so it holds fewer posts than the other platforms.
If those sources matter for your brand, a broader tool will see conversations a report here cannot. See the comparison pages for what each tool lists, with sources.
Not on every platform. X is the freshest: the live pulse on Agency reads X posts from the last minutes. Reddit and TikTok data can lag by hours. A tracked brand refreshes weekly on Pro and daily on Agency, so a tool that monitors many networks continuously may notice some changes sooner.
Sentiment is read in English, Spanish, Portuguese, Chinese, German, Arabic, Dutch and Swedish only; posts in any other language are left out of a report. Several listening tools state sentiment in dozens of languages or more, so for a brand talked about in other languages they will count posts a report here leaves out.
A new report reaches back up to 90 days. A tracked brand keeps its history and builds on it with every refresh, so its 180 and 365 day views open once its history covers them. Some enterprise tools sell archives that go back years, which a report here cannot reach.
English, Spanish, Portuguese, Chinese, German, Arabic, Dutch and Swedish. Each brand chooses its languages, any combination (English by default), when you add it or later on the brand; a language turned on later labels the posts already stored in it, with no report used. Every language is labeled with the same questions; posts in other languages are left out.
From the post itself or, on X, from the public profile location of an author we already have on record (new authors' profiles are not looked up for now). When neither names a country the post counts as unknown, and every country breakdown shows the share of posts with a known country. We never guess a country from the language.
Yes. Add a brand and choose Person or Topic: you can analyze a public figure or a public topic, with questions and default topics fitted to each. They are marked Beta and may change. Add a person by full name; it must be a public figure, and tracking or profiling private individuals is not allowed (see the terms).
The facts are measured in code from the report's own posts. With a change our significance test confirms, its two blocks of weeks are compared, and the change is split by topic, platform and intent into how much each group's share and sentiment moved. With no confirmed change, the narratives across the whole range are described, with no significance claimed. Our analysis model then writes a one-sentence takeaway, a summary and up to three findings from those facts and the posts, each a concern, a positive signal or something worth watching, marked as an observation, an interpretation the posts support or an unverified hypothesis, with the posts it cites, why it matters and what to look into next. Fewer findings, or none, when the evidence is weak; the tiles beside them are measured figures, never written by the model.
Checks run in code before anything is shown: every number must match a measured figure, every cited post must be one of the posts sent, and no conclusion may state a cause. A failed check gets one retry; if it fails again, the report shows the measured facts alone.
An explanation is written when the report finishes, and for a refresh only when the conversation changed materially. Each shows its version and analysis date; reports written before this method show their earlier measured shifts only, never as a cause.
Open Accuracy review from a brand's Topics and data page: answer relevance, sentiment, topic and intent for one post at a time before the label is shown, and see how often you agree. Where you disagree, your answer replaces the label in your own reports and the numbers update; you can also correct any single post from the explorer. Corrections spend no reports and never change anyone else's reports.
The live examples on the home page are real brands we track weekly, from real posts: the same report you get, not sample data. Anyone can open them read-only, without an account, at no cost.
Choose Ask a question in the menu, type the question, what it is about and the range (7, 30 or 90 days). We check the question first, which uses no report, then you confirm the claim, the search terms and the posts found for those dates; Run uses 1 report. The answer counts the posts that say the claim, reject it or discuss it, with shares weighted by platform and the posts behind every number. It describes what posts claim, never whether a claim is true, and with fewer than 200 posts about the claim the report is returned. Claim, comparison and discovery questions all run.
Three kinds. A claim question asks whether posts say something or reject it. A comparison asks which of two names the posts that compare them favor, with the mixed ones counted. A discovery question asks what people complain about: we propose up to seven themes from sample posts, you edit them on Confirm, and they are fixed when the question runs. Each answer shows counts first and shares weighted by platform, by platform and by week, findings and the posts behind every number; posts that are not about it, not a comparison or not a complaint are counted and never in a share. Every question uses 1 report.
Each report lists the questions people asked about the brand, grouped by topic and most engaged first, each with its platform, date and a link to the post, ready to answer. The section shows when a report has enough questions in its window.
Open the posts explorer from any report: filter by platform, sentiment, topic, intent, country, campaign, date or author, search the post text and sort. Every post shows its labels, and Pro and Agency export what you filtered as CSV. Live examples open it read only.
The 10 authors with the most engagement on posts about the brand in the window, each with its platform, posts, engagement and how its posts split between negative, neutral and positive. The brand's own accounts are left out. See posts opens the posts explorer on that author; the printable report lists them too. They come from the posts the report labeled, so on a sampled platform an author outside the sample is not listed.
Open Compare and pick 2 to 4 brands you have reports on, or choose competitors on a report. They show side by side from each one's latest report, on the same dates: sentiment over time, the tracker, topics and share of voice.
On Pro and Agency, add a campaign to a brand with its dates, tags and keywords. Its posts are measured against the brand's baseline before it, and after it ends you can print its impact report as a PDF: before and during, with intervals and the explanation.
Tracked brands (Pro: 3, Agency: 5) refresh on a schedule (weekly on Pro, daily on Agency) at no report cost once you start tracking them, which uses one report. Their campaigns update with each refresh; see how often brands and campaigns update.
Tracked brands and their campaigns update weekly on Pro and daily on Agency. On Agency, a live-mode day bought as an add-on updates X posts of one brand every 15 minutes for that day while its campaign runs. Every campaign gets a final update after it ends.
A campaign updates only while its brand is tracked. X is the freshest platform; data from the other platforms can lag by hours. Refreshes, live mode and the final update use no reports.
Yes. Pause in the brand's tracking settings: pausing keeps the history and frees the slot, and resuming needs a free slot again.
Stop tracking and Delete are on the brand's Edit page and in the menu (⋯) of its row on Your brands. Stopping ends the scheduled refreshes and alerts, frees the slot and keeps the history; tracking again uses 1 report. Deleting removes the brand with its reports and history, campaigns and share link, after you type its name; reports already used are not refunded. Only the account owner can delete a brand.
On Pro and Agency, create an alert rule for a tracked brand: a metric, a direction, a threshold and a window, optionally only when the change is significant. The rule shows how often it would have fired in the last three months. After each refresh we email you, at most once per rule a day, and call your webhooks if you have any. Switch a rule off to stop it.
On Pro and Agency, paste a Slack incoming webhook URL in Alerts and send a test message. Every alert that fires then also posts there with the brand, the change, its margin and a link to the report. If Slack refuses 3 posts in a row, the URL turns off and Alerts says so; paste it again to turn it back on.
On Agency, Pulse now fetches the last 12 hours of X posts for a tracked brand, minutes fresh, and labels them; other platforms show how fresh their data is. War room (live mode) does the same every 15 minutes during a campaign's dates. Both are included and spend no reports.
Live mode is part of Agency, and its days are bought as an add-on, $9 each, from your account page; Pro does not include live mode. One day covers one brand. A day counts once when live mode runs at any time that day (UTC), so switching it on and off again the same day uses one day. A bought day is kept until used, for 365 days.
The campaign page shows how many days are left this calendar month (UTC), and we email you once when one is left. With none left, live mode cannot start on a new day, and a running war room stops at midnight UTC. Days renew on the first day of each month. Live pulse, scheduled refreshes and reports are not affected.
On a campaign, choose Add creator posts and paste up to 10 post links from X, Instagram or TikTok. Each post shows its creator, date, reach (X impressions, Instagram and TikTok plays), likes, comments and shares, read when you add it, daily while the campaign is live and once when it ends. Pro includes 3 creator posts per billing period and Agency 5; adding them uses no report. More come as an add-on of 5 for $9, used within 365 days; each of those is followed for up to 30 days while your plan is active.
X replies and Instagram comments are read newest first and labeled like any post: the card shows their sentiment with its margin, the top themes and example comments, and says how many comments it read out of the post's total. Comments in English, Spanish, Portuguese, Chinese, German, Arabic, Dutch and Swedish are read, whatever languages the brand uses. If our data source stops answering partway, the card says how many comments were read. TikTok comments aren't available, so a TikTok post shows its figures and links to the conversation around the campaign instead.
If our data source doesn't have a TikTok post yet, its card says so and we check again daily for 3 days. A post that can't be read uses none of your creator posts and no report.
Every paid plan can download the sentiment chart as a PNG. Pro and Agency can also print a PDF report (US Letter or A4) and download CSV files of posts with their labels and of the numbers. Exports spend no reports. A CSV of posts includes public authors and links: keep it safe and do not resell it.
On Agency, set your company name and upload a logo in Account. They replace ours on your printed reports, campaign impact reports and every page your client logins see, with a "Prepared by" line if you want one. We store the logo only for that. Remove it anytime in Account; deleting your account deletes it too.
On Agency, create a key in Account, API keys; it is shown once, so copy it then. Send it as a bearer token. The API starts analyses, starts tracking and reads reports, numbers and posts, using the same reports as the app, and every answer states your reports left. Each key can make 60 requests a minute. Revoke a key in Account if it leaks. See the API documentation.
On Agency, add an https endpoint in Account. We send report.ready, report.failed and alert.fired, each signed with your endpoint's secret so you can check it came from us. A delivery that is not accepted is retried for about 8 hours. Remove the endpoint to stop deliveries. Details are in the API documentation.
You can delete your account yourself on the Account page: your brands, reports, campaigns, alert rules, API keys, webhook endpoints, uploaded logo and plan are deleted at once and any subscription is cancelled.
On Agency, invite up to 3 teammates from Account, Team. They sign in with their own email and share your reports, brands and tracked brands; billing, API keys, webhooks and the team stay with you. Remove a teammate anytime.
On Agency, give up to 10 clients a login from Account, Team, Client logins, each to the brands you choose. A client reads those brands' reports, posts and campaigns and nothing else, under your company's name and logo when you set them.
On Agency, create groups in Your brands and move each brand into one. The sidebar shows each group as a heading you can collapse; ungrouping keeps the brands.
A code for free reports goes in Account, Reports, Have a code? (once per account). A discount on Pro or Agency applies at checkout, or open the pricing link that came with it.
Share the link in Account, Refer a friend. When someone signs up through it and first pays, you both get 1 free report.
Yes. Account, Your data downloads every brand, report and the posts of each brand's latest report as CSV files in one zip. It spends no report.
Stored posts are not re-checked for deletion for now, so a post deleted after we stored it stays in your reports. To have a post taken down, write to support and it is left out of every report from then on. The status page shows whether everything is working.
Yes, choose Continue with Google on the sign-in page, or use the email link. Google shares your email address and name with us, never your password.
Write to support@senti.now. To have a post or an analysis taken down, write to the same address with its link.