Bluesky Starter Packs 2026: A Creator Playbook

What a Bluesky starter pack actually is

If you have spent any time on Bluesky in 2026, you have seen starter packs. They look like a curated grid of avatars — usually 20 to 50 accounts — that someone assembled around a topic. A single click on Follow All subscribes your account to every handle in the grid. The mechanism is older than Bluesky: Twitter Lists did this in 2009, and Mastodon lists did it again in 2018. What makes starter packs different is that the curator can publish the pack as a public URL, and Bluesky's Discover surface promotes new packs into the recommendation engine.

That last piece matters. A Twitter List is invisible unless someone shares it; a Bluesky starter pack is indexable. Once you publish one, it shows up in the Starter Packs tab under Discover → For You, and Bluesky's feed of starter packs is one of the three surfaces the app opens to by default. A well-titled pack in a tight niche can pull in 500 to 2,000 follows in a week — more than the median account gains in a year of organic posting. For a creator who has spent eighteen months building a niche audience on X and now finds the algorithm punishing them, the math is not subtle.

Starter packs are not magic. They are a curation primitive. What they solve is the cold-start problem: a new Bluesky user who lands in a niche pack gets a useful feed on day one, instead of an empty Following tab and a strong reason to leave. The pack does the work of "here are the fifty accounts worth reading about kubernetes", and the new user gets a feed without having to follow anyone manually.

How Bluesky's curation algorithm ranks starter packs

Three signals matter, and they are not equal in weight. The first is the follow-through rate: the share of people who click Follow All and stay subscribed for at least seven days. A pack where 80 percent of click-throughs unfollow within a week gets de-ranked fast, because Bluesky's quality team treats the metric as a proxy for relevance. The second signal is niche density: packs where every account overlaps the same topic — say, all the React core team members, or all the indie game developers shipping in 2026 — rank higher than packs that mix a React maintainer with three crypto influencers and a meme account. The third is curator authority: packs from accounts that already have a verified-blue check, or that are followed by a high-authority account, get a small boost.

The result is a curation ladder. New accounts building their first pack compete on follow-through and niche density. Established creators with verified handles compete on authority. A pack that clears all three signals moves into the Starter Packs tab and gets a steady trickle of organic discovery for weeks. A pack that misses on follow-through or density gets ignored within days.

Build a starter pack in five steps

The setup is genuinely short. You need a Bluesky account that is at least thirty days old (a soft anti-spam guard; new accounts can edit packs but cannot publish them) and a list of handles that actually belong together.

1. Pick the niche tighter than you think

The number one mistake is packing a niche too wide. "Tech people on Bluesky" is a list, not a pack, and it does not rank. "Indie Rust library maintainers shipping in 2026" is a pack. The narrower the topic, the higher the follow-through, and the higher the follow-through, the more Bluesky promotes the pack. Aim for 20 to 40 accounts in a coherent niche, never 200.

2. Audit each handle before you add it

Open the profile of every account you plan to add. Check that the bio still fits the niche, that the account has posted in the last thirty days, and that the last twenty posts do not include spam, abuse, or off-topic pivots. Starter packs that include dormant or off-topic accounts get the curator's whole pack de-ranked. Spend fifteen minutes auditing; it saves you from being quietly shadow-suppressed.

3. Use the official pack builder, not the API

The web UI is the safest place to build a pack. Go to Settings → Starter Packs → Create new. The API endpoint is also available, but it is rate-limited and a misformed payload returns a 422 with no useful error. Stick to the UI for the first version; once the pack is stable, you can automate via the API for monthly refreshes.

4. Title and short description carry half the rank

A pack titled Cool accounts is invisible. A pack titled Bluesky React Core 2026 is searchable in two different ways: as a feed candidate for new React-curious users, and as a query match for "react core bluesky" in the in-app search. Spend a full sentence on the description. Aim for the niche in plain English plus a date or year marker; the year marker tells new users the pack is current.

5. Publish, then pin the pack URL in your bio

Click Publish. The pack now has a public URL like https://bsky.app/starter-pack/your-handle/your-pack-id. Pin that URL in your profile bio, share it once on your existing X timeline, and let Bluesky's discovery surface take over from there. Do not reshare the pack URL every day — that is the spam pattern Bluesky's quality team watches for.

Five archival strategies that survive an account swap

Here is the part most guides skip. Starter packs disappear when their curator goes dormant, gets suspended, or rotates their handle. The accounts inside the pack keep working — the pack itself does not. If you have built a pack that your audience relies on, you owe them a backup. Five strategies, ranked by cost.

The fifth row is the realistic default, and it is why every creator who builds a pack they care about ends up rebuilding it from scratch six months later. Pick any of the first four. The Google Sheet is the cheapest; the API snapshot is the most rigorous; the ThreadGrab archive is the one that lets you actually read what was in the pack after it changes.

How starter packs compare with X Lists and Mastodon lists

X Lists and Mastodon lists have the same primitive — a curated set of accounts you can subscribe to in one click — but the discovery surface is what changes the math. An X List is invisible unless someone shares it; Mastodon lists are searchable within an instance but not federated; a Bluesky starter pack is part of the app's default navigation. For a creator trying to reach an audience that has not yet heard of them, Bluesky starter packs are the only one of the three that runs discovery for you.

There is also a moderation difference. X Lists inherit the list owner's suspension status silently — if the list owner gets banned, the list vanishes with no recovery. Mastodon lists are tied to the local instance, and instance defederation removes them. Bluesky starter packs are tied to the curator's DID, which is portable: if the curator changes handle, the pack URL keeps working as long as the DID is still active. That portability is what makes the archival strategies above viable: you can rebuild the curator identity without losing the pack definition.

Common mistakes creators make with their first pack

Three patterns show up in the first three months. The first is the all-stars pack: a creator adds the fifty biggest accounts in their field, follows the entire pack from a brand-new account, and expects discovery. The algorithm reads the follow-through as curator only, audience none, and ranks the pack at zero. The fix is to keep the curator's own feed active: post, reply, and re-share the pack over weeks, not once.

The second pattern is the stale pack: a pack that was built in 2024 and has not been refreshed since. Half the accounts have changed niche, a quarter have gone dormant, and a handful have rotated handles. The pack's follow-through collapses. Refresh the pack quarterly. Bluesky does not auto-update.

The third is the over-specific pack: a creator niches down to fewer than ten accounts because they want a tight list. The discovery engine refuses to promote a pack that small — the threshold is around fifteen accounts. Aim for twenty-five to forty.

What to watch over the next 30 days

Three things will probably land before mid-August. The first is paid starter packs: a creator-tier feature where a pack URL becomes a paywalled follow-gate, letting creators monetize the audience-building step. The second is cross-platform pack export: a Bluesky API endpoint that exports a pack as an OPML or JSON list, so the same curation can be republished on Mastodon or as an RSS bundle. The third is starter pack analytics for curators: per-pack click-through, retention, and unfollow-source graphs, surfaced inside the pack builder. All three would change the growth math again.

Until those land, the playbook is the one above: pick a tight niche, audit every handle, title the pack with the niche in plain English, publish, then archive the definition outside Bluesky's servers. Five steps for the pack, five strategies for the archive. The archive is the part creators skip and then regret six months later.

Frequently Asked Questions

What is a Bluesky starter pack in plain English?

A starter pack is a curated list of twenty to fifty Bluesky accounts that share a niche, published under a single URL with a Follow All button. Anyone who clicks the URL sees the avatars, can subscribe to every account in one click, and gets a useful feed from day one. Bluesky's Discover surface promotes starter packs the same way it promotes accounts, so a good pack pulls in organic follows for weeks after publication.

How many accounts should a starter pack include?

The sweet spot is twenty-five to forty. Fewer than fifteen gets ignored by the discovery engine. More than fifty dilutes the niche signal and tanks the follow-through rate, because new users are less likely to commit to fifty follow clicks than twenty-five. Audit every handle before you add it: dormant accounts, off-topic pivots, and spam are all reasons to leave a handle out.

Can I edit a starter pack after I publish it?

Yes, but the edit changes the pack's ranking. Removing accounts resets the follow-through metric. Adding accounts helps the niche density but resets the engagement clock. The safe pattern is to refresh the pack quarterly, in a single edit window, and not in dribs and drabs. Drafts are visible to you as the curator but not to the public until you publish.

What happens to my starter pack if my account gets suspended?

The pack URL stops resolving, and any new users following the pack's Follow All button get an error. Existing subscribers keep their follows; the pack definition itself is gone. This is the case for archiving the pack definition outside Bluesky's servers — a Google Sheet, a JSON snapshot via the API, or a ThreadGrab archive. The accounts inside the pack keep posting, but you lose the curation URL.

How does ThreadGrab fit into a starter pack workflow?

ThreadGrab's social-archive tools let you pull the recent posts from every handle in your pack into a single Markdown export, organized by account. That is the read-side archive: if your pack's curator handle ever rotates or your pack URL goes stale, you still have the recent output of every account you curated, indexed and searchable. Pair the pack definition (a JSON snapshot) with the ThreadGrab Markdown export, and you have a pack that survives any platform event.

Are starter packs the same as Twitter Lists?

The mechanic is similar — a curated set of accounts you can subscribe to in one click — but the discovery surface is different. X Lists are invisible unless shared; Bluesky starter packs are part of the app's default navigation and get organic discovery from Bluesky's recommendation engine. For a creator trying to grow an audience, starter packs are the only one of the two that runs discovery for you.