Last updated: November 2026
Instagram has capped posts at five hashtags. Its head, Adam Mosseri, says hashtags help search but do not increase reach. AI caption generators still earn their keep as drafting tools. A hashtag generator that outputs 30 tags is solving a problem the platform removed.
One App Store listing for a popular hashtag app advertises “top 10, top 20, top 30 hashtags” per post. Instagram announced on December 18, 2025 that posts and Reels would allow up to five, down from 30 (InstantDM, quoting the @creators announcement; Business Today).
That mismatch defines the category in late 2026. The tool that produces volume is no longer the useful tool. The useful one filters. It helps you choose three to five tags you can defend, and it writes a caption whose wording does the discovery work the tags used to do.
This guide is built to be checked, not just read. It sorts the evidence by strength and names the tools. It gives you prompts, a checklist, and a test protocol you can run on your own account. It also says plainly where nobody, including me, has the answer.
Scope: Instagram is the core. TikTok and LinkedIn get short, evidence-graded notes. X, YouTube and Pinterest are not covered.
Table of contents
- The short answer
- What has changed since the 5-tag announcement
- Where sources disagree
- How strong is the evidence that hashtags work?
- The Fanpage Karma dataset, read carefully
- What an AI caption and hashtag generator actually does
- Tool comparison for the 5-tag era
- The 3–5 tag selection framework
- Prompt library
- Verification checklist
- When to use AI and when to do it manually
- What a proper test would look like
- Myths, checked
- TikTok and LinkedIn: short notes
- What is still unknown
- What to do next
- FAQ
The short answer
- Do hashtags still matter on Instagram? Slightly, and mainly for search. Instagram’s own head says they do not increase reach. A large vendor study finds a small positive reach effect at five tags. Those two positions conflict, and neither is a controlled experiment.
- Do I need an AI hashtag generator? No. A free generator or a general AI assistant with a good prompt does the one job a generator can do: proposing candidates. The valuable step is verifying them.
- Is a paid hashtag app worth it? On current evidence, rarely. Pay for workflow (scheduling, approvals, analytics), not for text generation.
- What does the discovery work now? The words in your caption, including the first line, plus the three to five tags you can defend.
What has changed since the 5-tag announcement
Time-sensitive: recheck within six months.
| Then (before December 2025) | Now (late 2026) |
|---|---|
| Up to 30 hashtags per post or Reel | Five per post or Reel, per Instagram’s @creators announcement |
| Volume tools made sense: more tags, more shots | Volume tools are mismatched to the rule. Some app listings still advertise 20 to 30 tags per post |
| “Add a hashtag block” was standard advice | Mosseri’s position is that hashtags are for search, not reach |
| Hashtag choice was a numbers game | Hashtag choice is a filtering and verification problem |
| Caption was mostly a vehicle for tags | Caption wording carries more of the discovery load |
Rollout details still conflict across sources (see below). Treat five as the rule even if your account behaves differently.
Where sources disagree
Well supported:
- The cap exists. Instagram’s @creators account announced it, per InstantDM’s screenshot and quote. Several outlets reported it the next day.
- Mosseri’s position is consistent. He has said repeatedly that hashtags are for search and do not push content to more people (Mobile India; DesignRush).
Disputed or unsupported:
- Rollout. One report says the limit took effect immediately (The Mainstream). Another says it was rolling out gradually across accounts (SocialK). A third, Lilach Bullock’s site, states Instagram does not enforce a five-tag limit and that any cap comes from third-party tools or test groups. That contradicts Instagram’s own announcement. It may reflect accounts that had not yet received the change, but no source I found establishes that. If your account still accepts more than five, treat the cap as the rule anyway.
- Penalties. A hashtag-tool page claims posts above the cap are “silently demoted” from Explore and Reels, and that caption and first-comment hashtags carry identical algorithmic weight (SocialCal). It cites no primary source. Nothing from Instagram that I found supports either claim.
[Insert screenshot: the @creators announcement, with date]
How strong is the evidence that hashtags work?
Most hashtag advice rests on one of four kinds of evidence. They are not equally strong.
| Tier | Evidence | Example | How far to trust it |
|---|---|---|---|
| 1 | Platform statements | Instagram’s cap announcement; Mosseri’s remarks | Strong on what the platform says. Silent on how much hashtags help search |
| 2 | Large observational datasets | Fanpage Karma, 1.6M Instagram posts | Useful, but correlational and published by a company that sells a hashtag tool |
| 3 | Smaller vendor datasets | LinkedIn analyses (below) | Weak. They contradict each other |
| 4 | Unsourced assertions | “Internal TikTok guidance,” “silently demoted” | Treat as rumor until a primary source appears |
When you see a hashtag claim, place it on this ladder first. Most of the confident claims online sit on tier 3 or 4.
The Fanpage Karma dataset, read carefully
Fanpage Karma analyzed 1.6 million Instagram posts. Its findings:
- 45% of posts used no hashtags, 14% used five, and 7% used more than five.
- Posts with five hashtags had about 2% more reach (relative to follower count) than posts with none.
- Posts with more than five lost about 3%.
- Posts with one to three hashtags did slightly worse than posts with none.
- Engagement was highest at four to five tags, at +1% and +3% versus zero to one tag.
Why to hold these numbers loosely:
- It contradicts Mosseri. The study reaches the opposite conclusion from his statement, and the publisher says so.
- Conflict of interest. The publisher sells a free hashtag tool, which gives it a reason to find that hashtags matter. That does not make the data wrong, but it is a reason to want independent replication, and I found none.
- Missing methodology. The page, as I read it, does not state the data collection period or how accounts were selected.
- Selection effects. Accounts that use exactly five hashtags may differ systematically from accounts that use none, in ways unrelated to the hashtags. Careful, professional accounts may do both things.
- Correlation only. Nobody assigned hashtags at random.
- Reach is not search. The study measures reach and engagement. It does not measure search ranking or search impressions.
The effect size matters more than the direction. Two percent is small. In my judgment, a single account’s post-to-post reach varies by far more than that. A 2% hashtag effect would be undetectable in your own data unless you ran a very large number of posts.
Check this against your own Insights: compare your best and worst posts from the last month. If the gap is many times 2%, hashtag optimization cannot be your priority.
[Insert example Insights screenshot: best vs. worst post of the last 30 days, annotated with the gap]
What an AI caption and hashtag generator actually does
Most of these tools do not contain a hashtag database. They wrap a language model and, in the case of photo-captioning apps, an image-understanding model. The model writes plausible text. It does not look up what is popular.
[Insert workflow diagram: Input → Prompt → Model → Candidates → Your check, with the failure point under each stage]
| Stage | What happens | Typical failure |
|---|---|---|
| 1. Input | You supply a topic, photo or tone | Vague input gives generic output |
| 2. Prompt | The tool adds platform and length instructions | Platform not stated gives one-size copy |
| 3. Model | A text or vision model writes candidates | No live data, so “trending” is a guess |
| 4. Candidates | You get captions and tags | Stale, banned, or off-topic tags |
| 5. Your check | You verify before posting | A skipped check lets errors reach the post |
The failures are editorial analysis of how language-model tools work, not measurements of any specific product.
Stage 5 is the only stage that fixes the others. A tool that shows post counts and where they come from does part of that work for you. A tool that outputs a bare list of tags leaves all of it to you.
At stage 3, the model produces tags from patterns in its training data. It can write a convincing list of plausible tags, but it cannot know which are currently active, banned, or saturated unless the tool connects it to a data source. Apps advertise “live updates” and “trending” tags (example listing). I could not verify what data sits behind those claims, and a store listing is a first-party claim, not independent evidence.
Practical failure modes:
- Generic captions. A thin prompt yields interchangeable copy. The fix is input, not a better tool: give the model the specific detail that makes the post yours.
- Invented relevance. The model may attach tags that sound right but describe a different community.
- Platform mismatch. Caption norms differ by network. Vista Social’s generator page lists different character limits by platform, and ContentStudio notes that length, tone, and hashtag behavior all vary. A caption written for one network often reads wrong on another.
- Language coverage. One app listing supports captions in English, French, or Spanish only. Others advertise multiple languages. Check before you commit.
- Data handling. An app that captions photos must analyze the image somewhere. Whether that happens on your device or on a server varies by app, and the privacy policy is where to find out.
- Sparse topics. For a small local business or a narrow hobby, there is little for a model to draw on, and output drifts toward generic.
Tool comparison for the 5-tag era
How to read this table. The verdicts are my editorial judgment. They are based on what each tool type can do structurally and on the facts cited in this article, not on a controlled hands-on benchmark. I have not verified paid-tier prices or current feature sets for every tool. Where I have no verified figure, the table says so. Confirm pricing and features on the vendor’s page before you buy.
The criteria that matter now: Does it help you filter rather than add volume? Does it show where its data comes from? Is it honest about the five-tag limit? What does it do with your photos? What does it cost?
| Tool or type | What it is | Data source (what I can verify) | Cost (what I can verify) | Fit with the 5-tag reality | Verdict |
|---|---|---|---|---|---|
| General AI assistant (ChatGPT, Claude, Gemini) with a good prompt | Language model, no hashtag database | None live unless browsing is on. Tags are plausible guesses | Free tiers exist; paid plans vary, so check each vendor | Strong if you prompt for ten candidates and verify. Best at caption drafting | Free is enough. Best starting point. Use the prompts below |
| Ahrefs free hashtag generator | Free web generator | Not verified | Free per the vendor’s page | Good for candidate ideas. Verification is still on you | Free is enough |
| Vista Social free generator | Free generator with per-platform caption limits | Not verified | Free generator exists. Paid scheduler tiers not verified | Useful for platform-specific caption length | Free is enough. Paid only if you need the scheduler |
| ContentStudio free generator | Free generator, described by the vendor as no sign-up | Not verified | Free per the vendor | Fine for drafting; notes that tone and length vary by network | Free is enough |
| Fanpage Karma hashtag tool | Free tool from the publisher of the 1.6M-post study | Vendor’s own dataset | Free per the vendor | Its suggested mix (below) is a correlation from a vendor study | Free is enough. Use it, but remember the conflict of interest |
| Weekly-subscription caption and hashtag apps (for example, the App Store listing priced at $3.99 per week) | Consumer app wrapping a language or vision model | Claims “live updates” and “trending,” unverifiable from the listing | $3.99 × 52 = $207.48/year on the weekly plan. Other tiers’ billing periods aren’t stated in the listing text | Mismatched if it still sells 20 to 30 tags per post | Skip, unless it does something a free tool demonstrably cannot |
| Scheduler suites with AI drafting (Metricool, Vista Social, ContentStudio paid plans) | Scheduling, analytics, approvals, plus AI text | Varies. Metricool’s comparison notes that scheduling and hashtag support differ by tool | Varies; verify current plans | The value is workflow and analytics, not the AI text | Worth paying for only if you need the workflow. Not for the hashtags |
| Instagram’s own search | The platform’s tag and keyword search | The platform itself, and the only live source | Free | Essential. It is the verification step | Required. Use it on every candidate |
The asymmetric judgment: paying for a standalone caption-and-hashtag app is hard to justify on current evidence. A weekly subscription for a task a free tool does adequately is a poor trade unless the paid tool gives you something the free one cannot.
Prices and features are time-sensitive. The $3.99 weekly price was as listed in November 2026.
[Insert comparison table graphic for sharing]
The 3–5 tag selection framework
Pick by job, not by popularity. Each slot answers a different question.
| Slot | Job | Example for a post about a sourdough bake (illustrative) |
|---|---|---|
| 1. Category | Which broad area is this? | #baking |
| 2. Topic | What exactly is it about? | #sourdough |
| 3. Topic or format | A second angle or the format | #sourdoughbread |
| 4. Niche | Which small community would care? | A tag whose top posts match yours, found by searching |
| 5. Optional | Location, event or series | A local or recurring-series tag, only if it passes the checklist |
Rules of thumb:
- Fanpage Karma’s suggested mix of two to three large, one to two medium, and one niche tag is a correlation from a vendor study. Treat it as a hypothesis to test, not a rule.
- Fewer than five is fine. The study found one to three tags did slightly worse than none, but at that effect size I would not read much into it.
- Never fill slots to hit five. A weak tag is worse than an empty slot, because you have to defend each one.
- Rotate part of the set. Reusing an identical set on every post removes the only variation you could learn from.
Prompt library
Replace the bracketed parts. Always state the platform, topic, and audience, because that is where generic output comes from.
Prompt 1: Search-first caption opener
I'm posting on Instagram about [specific topic and detail only I know].
My audience is [who they are and what they'd type into search].
Write 5 versions of the first line of the caption. Each must read as a natural
sentence that includes a phrase someone would actually search for.
No hashtags. No emojis. Keep each under 125 characters.Prompt 2: Ten tag candidates, labeled
Platform: Instagram. Limit: 5 hashtags per post.
Topic: [topic]. Audience: [audience]. Location relevance: [yes/no, where].
Give me 10 hashtag candidates, not more. Label each broad, mid-size or niche,
and explain in one line what community would use it.
Say clearly that your size labels are guesses and that you cannot see live
post counts. Do not claim any tag is trending.Prompt 3: Red-team my tag set
Here are my 5 planned hashtags: [list].
Post topic: [topic]. For each tag, tell me the most likely way it could be
wrong: off-topic community, banned or flagged usage, saturation, or a
different meaning. Then tell me which one you'd drop first and why.
Be skeptical. You cannot see live data, so list what I must check by hand.Prompt 4: Caption in my voice, from my notes
Here are three captions I wrote that performed well: [paste].
Here are the facts of my new post: [specific details, a number, one
first-person observation].
Write one caption that matches my sentence length and tone. Do not add any
fact I did not give you. Mark anything you invented with [CHECK].Prompt 5: Rewrite for another platform
Rewrite this Instagram caption for [TikTok / LinkedIn].
Keep every fact the same. Adjust length, tone and keyword placement for that
platform. Suggest at most [3-5] hashtags and mark them as unverified.
Caption: [paste]Verification checklist
Copy this into your notes app. Run it on every tag before you post.
HASHTAG VERIFICATION CHECKLIST
For each candidate tag:
[ ] Opened the tag in Instagram search
[ ] Top results match my topic (not a different community)
[ ] Post volume is neither tiny nor enormous (I can say what size and why)
[ ] I can name who uses this tag
[ ] Not spam-heavy, adult or unrelated in the top and recent results
[ ] Not a duplicate of another tag in my set (e.g. singular vs. plural)
[ ] Spelling, capitalization and accents correct (use CamelCase for readability)
For the set:
[ ] 3 to 5 tags total, each with a defined job
[ ] Caption's first line contains a phrase someone would search
[ ] At least one tag changed from my last post
[ ] Logged the post in my test sheet (see the testing protocol)When to use AI and when to do it manually
START: I'm about to write a post.
│
├─ Do I have a specific detail, number or story only I know?
│ ├─ NO → Gather that first. AI output will be generic without it.
│ └─ YES → continue
│
├─ Am I drafting caption wording?
│ ├─ YES → Use AI for options. Edit for voice. Check every fact.
│ └─ NO → continue
│
├─ Am I choosing hashtags?
│ ├─ Use AI to propose up to 10 candidates.
│ └─ Verify each in Instagram search BY HAND. Never skip this.
│
├─ Is the topic niche, local or fast-changing?
│ ├─ YES → Do it manually. Search the tags yourself. The model has little to draw on.
│ └─ NO → AI candidates are a reasonable starting point.
│
├─ Does the post involve sensitive images or private details?
│ ├─ YES → Check the app's privacy policy, or write manually.
│ └─ NO → continue
│
└─ Am I paying for a tool?
├─ For text and hashtags alone → Don't. Free is enough.
└─ For scheduling, approvals, analytics → Consider it.[Insert decision tree graphic]
What a proper test would look like
A real answer to “do my hashtags help?” needs a design that holds up. Here is a protocol you can run on your own account. It is a reproducible method, not a result. I have not run it, and I am not reporting data from it.
The question
Do three to five verified tags change search-driven views or reach per follower compared with no tags or one tag, on comparable posts?
Setup
- Fix what you can control. Same account, same content type (all Reels or all carousels), same posting window (for example, the same time slot on different days), same caption style.
- Two arms.
- Arm A: three to five tags that passed your checklist.
- Arm B: no tags, or one tag. Choose before you start and don’t change it.
- Alternate, don’t batch. A, B, A, B. Running all A first and all B second confounds the test with seasonality, news cycles, and algorithm shifts.
- Sample size. Aim for at least 20 posts per arm, 40 in total. This is a rule of thumb rather than a power calculation. With this many posts, you can only detect a large effect, and a 2% effect is not detectable at this scale.
- Pre-commit your decision rule before you look at results. Write it down.
What to record for every post
| Field | Why |
|---|---|
| Date and time posted | Controls for time effects |
| Format (Reel, carousel, photo) | Formats have different baselines |
| Arm (A or B) | The treatment |
| Followers at posting | For reach per follower |
| Reach or views at 7 days | Same measurement window for every post |
| Views or reach from search/hashtags (per-post source breakdown) | The metric that tests the search claim |
| Saves, shares, comments | Secondary signals |
| Anything unusual (collab, trending audio, reshare by a large account) | Flag outliers |
How to measure search impressions specifically
Instagram’s per-post Insights shows where views or reach came from. Open a post, open its insights, and look for the traffic source breakdown, which may include entries such as profile, hashtags, Explore, or search. The labels and available breakdown change between app versions and account types, so check your own view. If your app version does not break out search or hashtag sources separately, say so in your notes rather than assuming. In that case, you can only test total reach, which is a weaker test of the “hashtags help search” claim.
[Insert example Insights screenshot: per-post source breakdown, with the search/hashtag line highlighted]
Analysis you can do in a spreadsheet
- Compute reach per follower for each post.
- Compare the median of Arm A with the median of Arm B. Medians resist one viral outlier better than averages do.
- Look at the spread: put the posts on a simple dot plot by arm. If the two groups overlap heavily, you have no detectable effect.
- Apply your pre-committed rule. Example: “If the Arm A median is not above the Arm B median by more than the typical post-to-post spread, I will stop tuning tags.”
- Decide what the result means for effort. If you find no detectable difference, spend the time on the first line of the caption, keywords, and content quality.
What this test cannot tell you
- It cannot isolate hashtags from the content of the post, even with alternating. Different posts have different appeal.
- It cannot see Instagram’s ranking internals.
- With a small sample, “no difference found” does not mean “no difference exists.” It means the effect, if any, is too small to see in your data.
- It tells you about your account. It does not settle the platform-wide question.
Myths, checked
| Claim | What the evidence supports |
|---|---|
| “Posts above the cap are silently demoted” | One hashtag-tool page claims this and cites no primary source. Nothing from Instagram that I found supports it. Unsupported |
| “First-comment hashtags work the same as caption hashtags” | The same tool page claims this with no primary source. I found nothing from Instagram on the question. Unsupported, and also unrefuted |
| “30 hashtags gets more reach” | Contradicted by Instagram’s cap and by the Fanpage Karma data, where more than five performed about 3% worse. Outdated |
| “More hashtags = more reach” | Mosseri says hashtags do not increase reach. The vendor study finds a small positive effect at five. Neither supports “more is better.” False as stated |
| “AI tools know which hashtags are trending” | Language models do not have a live hashtag database unless the tool connects to one. Claims of “live” data are first-party and unverified. Unverified |
| “TikTok officially recommends three to five hashtags” | No primary source found. One page says TikTok has never published an official optimum. No primary source |
| “LinkedIn hashtags lift reach by a known percentage” | Three datasets point in opposite directions. Contradicted |
| “The five-tag cap is not enforced” | One site says so, which contradicts Instagram’s announcement. Disputed. Treat five as the rule |
TikTok and LinkedIn: short notes
TikTok
- The “3 to 5” rule has no primary source. Dozens of pages repeat it. One states plainly that TikTok has never published an official optimum (TTCalculator). Others cite “internal guidance shared at conferences,” which no reader can check. The number looks like consensus, but every page appears to cite the others.
- What multiple sources agree on: TikTok search reads caption text, so keywords in the caption carry weight alongside tags (SocialKit). That is also a secondary claim, but it is consistent across sources and with how the platform has been described.
- Practical takeaway: write the caption with search phrases in it. Treat tags as a supporting signal.
Three datasets point in opposite directions:
- One reports that posts with at least one hashtag got 85% more impressions, while median reach fell from 874 with no hashtag to 492 with five (WaveGen).
- Another reports posts without hashtags averaging 6,632 impressions versus 2,269 with them (Ordinal).
- A third cites an analysis of 10,000 posts finding about 12.6% more engagement and roughly 9% more reach for posts with at least one hashtag (ViralBrain).
These cannot all describe the same effect. The likely explanation is selection: people with large followings often skip hashtags, so no-hashtag posts look strong for reasons that have nothing to do with hashtags. That is an inference, not something these sources demonstrate. What they do show is that anyone quoting a precise hashtag lift on LinkedIn is quoting a number that a different dataset contradicts.
On guidance, sources also split. Some say LinkedIn recommends three to five hashtags, while another states LinkedIn has no official guidance on the number (DemandBird). I found no primary LinkedIn document either way.
Practical takeaway: run the same testing protocol on your own LinkedIn posts if it matters to you.
What is still unknown
- How much hashtags help Instagram search. Instagram has not published a figure, and no source I found quantifies it. Fanpage Karma’s data measures reach and engagement, not search ranking.
- Whether the cap counts first-comment hashtags the same way as caption hashtags. I found no primary statement.
- Whether exceeding the cap is penalized. No primary evidence.
- Whether the Fanpage Karma pattern replicates outside the publisher’s own data, and over what period it was collected.
- Whether AI-suggested tags perform differently from hand-picked tags. I found no controlled comparison.
- How any of this will change. Platform rules move. See the freshness note below.
Resolving the search question would take a controlled test on identical posts that differ only in tags, measuring impressions from search specifically. Instagram’s Insights shows traffic sources per post, so an account owner can run a small version of this (see the protocol above). A platform-level answer needs Instagram’s own data.
What to do next
Today (15 minutes):
- Open your last 30 days of Insights. Note your best and worst post. Compute the gap. If it is many times 2%, hashtags are not your biggest lever.
- Pick the one post you’ll publish next and write its first line as a search phrase (Prompt 1).
This week (three posts):
- Run Prompt 2 in a free generator or general assistant for ten candidates, not thirty.
- Run the verification checklist on each candidate in Instagram search. Note how many fail.
- Pick three to five by job using the framework. Rotate at least one tag per post.
Decide on tools:
- If most of the suggested tags fail the check, the tool is producing noise, and paying for it will not fix that. Stay on free tools.
- If you are weighing a paid plan, ask one question: does it give me workflow (scheduling, approvals, analytics) that I would otherwise do manually? If not, skip it.
This month (40 posts is the target, 20 per arm):
- Set up the test sheet from the protocol, with a pre-committed decision rule. Alternate arms. Don’t peek early.
- If the groups overlap heavily, stop tuning tags and put the time into hooks and keywords.
Keep this page honest: bookmark it. Prices, platform limits, and rollout status are as of November 2026. Recheck the Instagram cap rollout, LinkedIn hashtag guidance, and any quoted subscription prices within six months.
FAQ
How many hashtags can you use on Instagram now?
Five per post or Reel. Instagram announced on December 18, 2025, that the limit would drop from 30 to five. Rollout reports vary, and one source disputes enforcement, so treat five as the rule even if your account still accepts more.
Do hashtags still increase reach on Instagram?
Instagram head Adam Mosseri says hashtags help search and do not increase reach. A vendor study of 1.6 million posts found about 2% more reach at five tags versus none. It is correlational, and the publisher sells a hashtag tool. The effect, if real, is small.
Do hashtags help Instagram search?
That is the platform’s stated purpose for them. No source I found quantifies how much. The protocol above lets you measure it on your own account if your Insights shows a search or hashtag source breakdown.
Should I put hashtags in the caption or the first comment?
I found no primary source from Instagram that says. One tool’s page claims they carry identical weight but cites nothing. Since caption text is read for search, putting keywords in the caption is the safer bet. If you want evidence for your own account, test it with the protocol.
Are posts with more than five hashtags demoted?
Unverified. A hashtag-tool page says “silently demoted” and cites no primary source. Nothing from Instagram that I found supports it. In the Fanpage Karma data, posts with more than five lost about 3% reach, but that is correlational.
Is an AI hashtag generator worth paying for?
Usually not. These tools mostly wrap a language model with no live hashtag data. Free generators from Ahrefs, Vista Social, and ContentStudio exist, and a general assistant with a good prompt does the same drafting. Pay for workflow, such as scheduling and analytics, not for text generation.
How do AI hashtag generators work?
Most use a language model, and photo apps add an image-understanding model. The model writes plausible tags from patterns in training data. Unless the tool connects to a live data source, it cannot know which tags are currently active, saturated, or banned.
How do I know a hashtag is any good?
Open it in Instagram search. Check that the top results match your topic, that the post volume is neither tiny nor enormous, and that you can say who uses it. Use the checklist above.
What is the best number of hashtags: 3, 4 or 5?
No source proves an optimum. The Fanpage Karma data showed the strongest engagement at four to five tags, but that is correlational. Pick the number of tags you can defend, up to five.
Do hashtags work the same on TikTok?
Not necessarily. The popular “3 to 5” TikTok rule has no primary source, and one page says TikTok never published an optimum. Several sources agree that TikTok search reads caption text, so keywords in the caption matter.
Do hashtags work on LinkedIn?
Evidence conflicts. Three datasets give opposite results, and sources disagree on whether LinkedIn offers any official guidance. Anyone quoting a precise lift is quoting one dataset against others.
How much does a hashtag app cost?
One App Store listing shows a weekly plan at $3.99, which annualizes to $207.48 (3.99 × 52). Other tiers on the same listing do not state their billing periods in the text. Prices change often. Check the subscription screen before you buy.
Do AI-written captions perform worse than human ones?
I found no controlled comparison. If AI-drafted captions underperform yours, the likely cause is thin input. Add a specific detail, a number, or a first-person observation that only you could supply before changing tools.
Should I use the same hashtags on every post?
Rotating part of the set gives you something to learn from. Reusing an identical set on every post removes the only variation you could measure.
What should I do if my account still allows more than five hashtags?
Use five or fewer. Instagram’s announcement is the primary source, and nothing I found suggests that going over helps.
Prices, platform limits, and rollout status are as of November 2026. Time-sensitive claims in this article: the Instagram cap rollout, LinkedIn hashtag guidance, TikTok guidance, tool pricing and feature claims, and the per-post Insights breakdown labels.