Twenty-five straight answers

You give it a keyword, a platform, a language, a tone and a group size. The engine breaks your keyword into word parts, then combines those parts with thirteen intent-specific vocabulary banks — popular, trending, low competition, niche, viral, location, seasonal, branded, community, event, industry and the two competition tiers in between. It also folds in the shapes each platform genuinely rewards, so an Instagram set and a LinkedIn set for the same keyword come out looking different.

Nothing is pulled from a fixed list of finished hashtags. Every tag is built around the keyword you typed, which is why two similar niches produce genuinely different results.

It is free, and there is no account, no email wall and no trial that expires. You can run the generator as many times as you like, copy everything, and download the results as TXT or CSV without ever identifying yourself.

That is an estimated difficulty score from 3 to 99. Low numbers mean a smaller, more winnable pool of competing posts; high numbers mean the tag is crowded and your post is unlikely to stay visible for long.

The score is a heuristic. It is calculated from the tag's length, how many words it contains and which intent category it belongs to — not from a live feed of platform data. It is a way to rank the options against each other, not a measurement of any platform's index.

Because an exact number would be a fiction. No public API gives accurate, current post counts across all the major platforms, and any tool showing you "1,482,993 posts" for an arbitrary tag is either using stale scraped data or making it up.

Ranges are honest about the precision that actually exists. Knowing a tag sits in the "under 50K" band rather than the "5M+" band is the decision-useful part.

It depends heavily on the platform. Instagram allows thirty but around eight to fifteen relevant tags typically outperforms a maxed-out block. TikTok captions are short, so three to six specific tags beat a wall of generic ones. YouTube shows only the first three above the title. Facebook and X reward real restraint — two or three, and one or two respectively. LinkedIn sits around three to five. Threads attaches a single topic tag.

The generator shows the recommended number for whichever platform you have selected, and the "recommended mix" panel assembles exactly that many for you.

Copying all 100-odd generated tags into a post is the single most common mistake creators make. The mix takes a platform-appropriate number of tags and weights them toward low and medium competition, with only a couple of high-competition tags included.

That shape — mostly winnable tags, a few reach tags — is what gives a post a realistic chance of ranking somewhere rather than disappearing everywhere.

They work differently than they did five years ago. On every major platform, recommendation systems now lean more heavily on watch behaviour, caption text, audio and on-screen text than on tags alone. Hashtags have shifted from being a reach lever to being a categorisation and search signal.

That is still valuable. Tags tell the platform what your post is about, they make you findable in tag and keyword search, and on Pinterest and YouTube they feed directly into a search-driven discovery model. What no longer works is stacking thirty popular tags and expecting reach.

Testing across large samples has repeatedly found no meaningful reach difference between the two on Instagram. The choice is aesthetic: caption placement is simpler and keeps everything in one place, first-comment placement keeps the caption clean.

Pick one and stay consistent so you can actually read your own analytics.

Yes, though the lists circulating online are usually years out of date. Platforms restrict tags when they attract policy-violating content, and a restricted tag can quietly limit the reach of the whole post carrying it.

The practical defence is to check any unfamiliar tag by searching it on the platform before you use it. If the tag page shows a warning, shows no recent posts, or shows nothing at all, leave it out. This generator does not produce tags from known-restricted vocabulary, but platform lists change without notice, so a quick check is always worth it.

Yes. Selecting Spanish, French, German, Portuguese, Italian, Dutch, Indonesian, Turkish or Romanised Hindi swaps the connective vocabulary the engine uses when building compound tags, so you get tags a native speaker would plausibly write rather than English tags with a foreign keyword bolted on.

Your keyword itself is used as typed. If you want a fully localised set, type the keyword in the target language too.

Tone controls the adjective bank used to build compound tags. Professional pulls in words like expert, strategy and benchmark. Playful pulls in cosy, fun and chaos. Luxury reaches for signature, bespoke and refined. Minimal favours quiet, slow and essential.

It is the quickest way to make a generated set sound like your brand rather than like a tool.

By design. The generator is deterministic — the same keyword, platform, language and tone always produce the same tags, so a result you liked is reproducible and shareable via its URL.

Use Shuffle to advance to the next variation, or Regenerate to jump to a completely different arrangement.

Yes. Hashtags are short functional strings, not creative works, and nothing generated here is licensed or restricted. Use them for client work, paid campaigns, agency deliverables or your own accounts.

The one thing to check yourself is branded tags: if the generator suggests something that happens to match an existing trademark, that is coincidence, and using another company's brand tag to promote your own products can create legal exposure.

No. Generation happens per request and nothing about your input is written to a database or associated with you. There is no account system, so there is nothing to associate it with. See the privacy policy for the full detail on what the site does and does not collect.

Because that is how hashtags work — no spaces, no punctuation. A three-word keyword becomes one string, and compound tags built from it can look dense at first glance.

If a tag reads badly, skip it. Readability matters: tags that are hard to parse get typed less often, which means they get searched less often.

Guides are reference material — evergreen explanations of how a mechanic works, organised so you can come back and look something up. The blog is shorter and more current: platform changes, tactics, experiments and commentary.

Rotate meaningfully every few weeks rather than every post. Using an identical block on every post gives the platform less information about what each individual post is about, and it makes your own analytics unreadable because every post carries the same variables.

Keep a small core of tags that genuinely describe your niche, and rotate the rest based on the specific post.

Copy-pasting an identical thirty-tag block onto every post is a pattern associated with automation, and it is a reasonable thing to avoid. There is no confirmed penalty that platforms have documented publicly, but the behaviour is a poor fit with how ranking systems interpret content anyway.

Vary your tags because it produces better categorisation, not out of fear of a penalty nobody can verify.

Instagram, TikTok, YouTube, Facebook, X (Twitter), LinkedIn, Pinterest and Threads. Each one has its own tag ceiling, recommended count, native tag shapes and cultural staples baked into the engine.

Yes — TXT for pasting into a caption draft, and CSV for opening in a spreadsheet or importing into a scheduling tool. The CSV includes the category, difficulty score, tier and estimated pool band for each tag, which makes it easy to build and maintain your own tag library.

No, and the trending page says so directly. The lists there are editorially maintained patterns — the tag families that consistently carry volume in each niche — not a real-time feed. Anyone claiming a live cross-platform trending feed is overstating what public APIs actually make available.

For genuinely live trends, check the platform's own discovery surface: TikTok's Creative Center, X's trending panel, Pinterest Trends and YouTube's trending tab.

In Reels, tags work much as they do on feed posts, with the caveat that Reels captions are short and a handful of specific tags reads better than a block. In Stories, a tag can add your story to that tag's story reel, but the reach from it is usually small — one relevant tag is plenty.

Platform limits vary but are generous — long tags are technically fine. The practical limit is human: past about twenty characters, a tag becomes hard to read and nobody types it into search. The generator caps tags at twenty-eight characters and favours shorter, readable combinations.

Yes. There is no per-seat licence, no attribution requirement and no restriction on commercial use. Plenty of the people using this are running accounts for other people.

The CSV export exists specifically so agency teams can drop tag sets into shared documents and scheduling tools.

The tool is deliberately cheap to run — it is a small PHP application with no database, no third-party APIs and no per-request cost. That means it does not need to be monetised aggressively to survive.

If that ever changes, it will be stated plainly on the about page rather than quietly introduced. What will not happen is the pattern common in this category: a free tool that gradually becomes a paywall with a demo attached.