Second Life Bots · October 11, 2026 · 9 min read

Are Bots Allowed in Second Life? The Rules in 2026

Are bots allowed in Second Life? Yes. Linden Lab permits automated avatars, as long as the account is flagged as a Scripted Agent and the bot follows the rules that come with that status: a daily message cap, a short list of banned behaviours, and respect for estates that choose to keep bots out. That is the whole answer. The rest of this page explains each rule in plain language, what happens when a bot ignores them, and the checklist we give our own users. We sell bot features, so we have every reason to be precise here: a compliant bot is the only kind that lasts.

If you are new to bots in general, start with our complete guide to Second Life bots, which covers what they do and how the platforms differ. This page is the rulebook.

What counts as a bot: the Second Life Scripted Agent

Linden Lab's term for a bot is a Scripted Agent: an avatar account driven by software rather than by a person at a viewer. The policy says that accounts mainly operating as scripted agents must be identified as such. In practice, that covers the familiar SL bot types:

• Greeters that welcome visitors to a store or venue
• Group managers that send invites and notices
• Store and model avatars that stay logged in around the clock
• NPCs in roleplay regions
• Second Life AI bots that answer chat and IMs on their own

The test is not the technology. It is whether the account acts without a human deciding each action. A person typing in a browser viewer is not a scripted agent. The same account replying to visitors automatically at 3am is.

Are bots allowed in Second Life? What the policy actually says

The authority is Linden Lab's Bot policy on the Second Life wiki, together with Linden Lab's support FAQ on scripted agents and estates. Here is what it requires, rule by rule.

1. Flag the account

Any account that mainly operates as a scripted agent must be flagged. You do this on the Second Life website, at secondlife.com/my/account/sisa.php, while logged in as the bot account. It takes a minute. The flag is what lets estate tools, region traffic, and Linden Lab's own systems tell your bot apart from a person.

2. Stay under 5,000 messages per day

A user may send at most 5,000 messages per calendar day across all the agents they run. Two details matter:

• "Messages" is broad. It includes local chat, IMs, inventory offers, group notices, and invites.
• The cap is per user, not per bot. Run five bots and they share one 5,000 budget between them.

That second point catches people out. A fleet of greeters that each look modest can add up to a number the policy does not allow.

3. Avoid the prohibited behaviours

The policy names specific things a scripted agent must not do:

• Inflate parcel traffic. Parking bots on land to push up its traffic score is out.
• Put unreasonable load on regions or Linden Lab services.
• Automate purchases of Linden-owned mainland.
• Circumvent other residents' privacy settings.

Everything else in the Community Standards and Terms of Service still applies. A bot that harasses, spams, or disturbs people breaks the same rules a person would.

4. Respect estates that deny bots

Since March 2023, estate owners have had a deny_bots setting. It works estate-wide only (not per parcel) and blocks identified bots from the whole estate, except avatars on the estate's allow list and the estate managers.

Here is where the flag matters most. A flagged bot simply cannot enter a no-bots estate, which is the system working as intended. An unflagged bot that wanders into one is a policy violation, and estate owners can report it as Disturbing the Peace.

5. Expect consequences for breaking the rules

Penalties for violating the policy run up to suspension or termination of the account. Since the bot account is often the one holding your store's inventory, group roles, or customer history, that is a real cost.

Has anything changed recently?

We checked in October 2026 and found no policy change in 2025 or 2026; the wiki page was last edited in April 2023. Linden Lab has said it plans to update its policy on NPCs and agents, so watch for that.

This is a plain-language summary, not legal advice. Linden Lab's policy page is the authority, and it can change. Read it yourself before you run a bot.

Do Second Life AI bots follow different rules?

No. There is no separate category for AI. A Second Life AI bot that answers IMs, chats in local, or greets arrivals without a person approving each line is automation, so it is a scripted agent and the same five rules apply.

AI does change two practical things:

• Volume. A model can reply to everything, instantly, all day. That is how you burn through a 5,000 message budget without noticing.
• Responsibility. You own what the bot says. A model that improvises a wrong price or an off-tone reply is your message, sent from your account.

Our overview of AI bots for Second Life covers how to set one up so visitors find it useful rather than noisy. The short version: exact answers for known questions, AI for the rest, and hard caps on how much it can send.

A compliance checklist for any SL bot

Use this whether you build your own, rent from a hosted service, or use us.

1. Create a dedicated account for the bot. Do not automate your main avatar.
2. Flag it as a Scripted Agent at secondlife.com/my/account/sisa.php.
3. Add up your daily messages across every bot you run. Chat, IMs, offers, notices and invites all count toward one 5,000 cap.
4. Set per-person cooldowns so nobody gets the same greeting or offer twice in a short window.
5. Never park bots to inflate traffic, and do not use them to get around anyone's privacy settings.
6. Ask before placing a bot on someone else's estate. If it uses deny_bots, you need to be on the allow list.
7. Keep a log of what the bot did, so you can answer for it if someone complains.
8. Keep a way to switch it off fast, for when something misbehaves.

How Thunder Viewer lines up with the rules

We build the bot layer into a browser client, so we get asked about policy constantly. Our answer is the same every time: if your avatar runs unattended, run it on a dedicated account and flag that account as a Scripted Agent. We say this in our own docs and in every bot post on this blog, including the SmartBots alternatives roundup, because a hosted service and a browser client face the same rules.

The product is designed around the checklist above, not just the copy.

One outreach avatar per account

Arrival IMs, departure IMs, and region broadcasts reach strangers. If you link several avatars, only one of them is allowed to do outreach. Turning it on for a second avatar moves it there rather than adding a second sender. Several avatars greeting the same visitor is spam, and a message that has been sent cannot be unsent.

Cooldowns that claim the window before sending

This one we learned the hard way. Early on, a visitor received the same outreach message three times in two minutes. Their presence flapped (left, reappeared, left again as the region updated), and each re-trigger checked "have we messaged this person recently?" before the first send had recorded itself. Every check passed.

The fix was to make the cooldown claim the window atomically before sending, so the second and third triggers find it already taken. Greetings, arrival IMs and broadcasts all go through that claim now: repeat visitors get an outreach IM at most once per 12 hours, and a broadcast never reaches the same person twice in 24 hours. It is a small change in code and a large one in how a region feels.

Daily caps per account, not per avatar

Linden Lab counts messages across every agent a user runs. We count the same way. The daily allowance for AI replies and for marketing IMs belongs to your account as a whole, so linking a fifth avatar does not quietly multiply what your fleet sends. Broadcasts are throttled with a pause between sends and count against the same daily marketing cap.

An off switch at two levels

Every bot has a master on/off switch above its individual features, and an activity log listing everything it did: replies, greetings, marketing IMs and moderation actions. Above that, we can pause every AI bot on the platform in real time if something goes wrong, without changing anyone's own settings. When the pause lifts, each bot returns to exactly the state its owner left it in.

None of this replaces your own judgement. The caps keep you inside sane limits; the flag, the estate etiquette, and what your bot actually says are still yours. If you plan to run it around the clock, our guide to keeping a Second Life avatar online covers that side, and the Discord bot guide covers mirroring its activity to your team.

FAQ

Are bots allowed in Second Life?

Yes. Linden Lab allows bots under its Bot policy, provided the account is flagged as a Scripted Agent, stays within the 5,000 messages per day cap, avoids the prohibited behaviours, and respects estates that deny bots.

What is a Second Life Scripted Agent?

It is Linden Lab's name for an avatar account operated by software rather than a person. Accounts that mainly operate this way must be flagged at secondlife.com/my/account/sisa.php.

Do AI chat bots count as scripted agents?

Yes, if they act unattended. An AI that replies to IMs or local chat without a person approving each message is automation, so the account should be flagged and the same rules apply.

Can I run a bot on my main account?

We recommend against it. The policy requires accounts that mainly operate as scripted agents to be flagged, and a flagged account is blocked from estates that deny bots. Keep your main avatar for yourself and give the bot its own account.

How many messages can a Second Life bot send?

Up to 5,000 per calendar day, and that limit covers all the agents one user runs combined. Local chat, IMs, inventory offers, group notices and invites all count.

Can estates ban bots?

Yes. Since March 2023, estate owners can enable deny_bots, which blocks identified bots from the entire estate. Avatars on the estate's allow list and the estate managers are exempt. It cannot be set per parcel.

What happens if my bot is not flagged?

It is out of compliance with the policy. If it enters an estate that denies bots, that is a violation the estate owner can report as Disturbing the Peace. Penalties for policy violations go up to suspension or termination of the account.

Where is the official Second Life bot policy?

On the Second Life wiki: Linden Lab Official: Bot policy. Check it directly before you deploy, since Linden Lab has said it intends to update its policy on NPCs and agents.

Want a bot that is built around these rules from the start? Try Thunder Viewer free.