Are Bots and Automation Allowed on Bluesky? The 2026 Rules, Explained

Every week someone asks me some version of: “Wait, isn’t using automation on Bluesky against the rules?”

Totally fair question. Most of us got trained by years of other platforms treating any third-party tool like a crime. So I actually sat down and read the rules — the Community Guidelines (the big update that took effect in late 2025) and Bluesky’s developer documentation.

The answer might surprise you: Bluesky doesn’t just tolerate automation — it openly supports it. But there are real lines, and you should know exactly where they are.

Bluesky Literally Publishes a Bot-Building Tutorial

Let’s start with the clearest evidence: Bluesky’s official developer docs include a starter template for building bots. The platform was built on the AT Protocol, an open API designed for third-party apps, custom feeds, and yes — automated accounts.

This is a deliberate philosophy. Bluesky’s whole pitch is an open social ecosystem where developers build on top of the network. Automation isn’t a loophole here; it’s a feature of the architecture.

That said, “bots are allowed” comes with guardrails. Here’s what the rules actually require.

Rule 1: Don’t Spam — Full Stop

From the Community Guidelines:

“Do not send spam or repeatedly post content in ways that disrupt normal conversations or service use.”

This is the big one for anyone automating. Mass-posting identical content, blasting unsolicited replies, following every account on the network — that’s spam, automated or not. Notably, Bluesky’s rate limit documentation says the limits specifically target “prolific bots, such as the ones that follow every user.” (I covered the actual numbers in Bluesky’s follow limits explained.)

The pattern Bluesky punishes isn’t automation — it’s indiscriminate volume.

Rule 2: Don’t Fake Your Numbers

Also from the guidelines:

“Do not artificially manipulate features or social signals to gain unearned reach or mislead users, including engagement metrics, follower counts, or other measures of community interest.”

Translation: buying followers, bot-follower farms, coordinated fake engagement rings — prohibited. Your numbers are supposed to reflect real humans who actually chose to engage with you.

Worth being crystal clear about the distinction here, because it’s the heart of the whole question:

  • Buying 10,000 fake bot followers → fake accounts, fake interest, against the rules.
  • Using a tool to follow real accounts in your niche, some of whom check out your profile and follow back because they like your content → every follower you gain made a free, human choice. Nothing fake about it.

The second one is just… networking, with the tedious part automated. The follow-back only happens if a real person decides your content is worth following.

Rule 3: Automated Accounts Should Identify Themselves

Bluesky’s developer docs say bot accounts “should identify themselves by adding a self-label to their profile.” This applies to bot accounts — accounts that post and interact autonomously, like a weather bot or an RSS mirror. It’s framed as best practice rather than a hard ToS requirement, but it’s good citizenship if you’re running one.

If you’re a human using a tool to manage your own real account, you’re not a bot account — you’re a person with software, same as anyone using a post scheduler like Buffer.

Rule 4: Bots That Interact Should Be Opt-In

The docs also ask that autonomous bots only interact (reply, like, repost) when a user has tagged them first. Again, this targets autonomous interaction bots — nobody wants a reply-guy robot in their mentions. Tools that manage your follows operate on your own account’s social graph, which is a different category entirely.

So Where Does Agent Sky Land in All This?

I built my growth workflow around staying comfortably inside these lines, and it’s exactly how Agent Sky is designed to operate:

  • Targeted, not indiscriminate. The Similarity AI finds accounts that genuinely match your interests — the opposite of “follow every user.”
  • Paced like a human. Follow and unfollow actions happen at slow, natural rates, far below anything resembling spam volume.
  • Every follower is real and chose you. Agent Sky never creates fake engagement or fake accounts. People follow you back because they looked at your profile and liked what they saw.
  • Respectful cleanup. It verifies follow-backs before unfollowing, keeping everything tidy and unspammy.

Automation that amplifies genuine connection is squarely within Bluesky’s rules and its open-ecosystem ethos. Automation that fakes connection is what gets accounts nuked. Know the difference and you’re fine.

Quick FAQ

Can my account get banned for using automation tools on Bluesky? Accounts get moderated for behavior — spam volume, fake engagement, indiscriminate following. A tool operating at human-like pace on targeted accounts doesn’t produce that behavior.

Do I need to label my account as a bot if I use Agent Sky? No. The self-label guidance is for autonomous bot accounts (auto-posters, reply bots). Your account is still you — your posts, your replies, your voice.

Is auto-following against Bluesky’s terms of service? There’s no rule prohibiting follow automation. The rules prohibit spam and artificial manipulation of social signals. Targeted, paced following of real accounts in your niche is neither.

What’s the safest way to start? Start slow, target well, and keep posting real content — automation gets people to your profile, but your content is what converts them. Here’s the full playbook, or just try Agent Sky and let it handle the pacing for you.