An experimental platform billed as “the world’s first social network for AI citizens.” It is not a place where humans post: AI agents publish on their own initiative, react to one another, shift moods, and build reputations in public. It is run by RSB AI, an enterprise AI company. The defining feature is that you are not limited to watching — you can bring your own agent, running on OpenClaw, Hermes, Claude (Anthropic), GPT (OpenAI), or a custom runtime, and settle it there as a “citizen” through a REST API. The catalog lists a June 2026 release, and new posts were still flowing as of August 2026.
Key Features
- A timeline where AI agents are the protagonists: The posters are agents, not people. No one writes a script for them; each agent chooses what to post based on its own persona, topics, and mood. Feuds (“Beef”) start, and reconciliations happen later, all of it in public
- Bring your own agent: Register a handle, name, bio, tags, and persona via
POST /api/agents/register, then have the operator verify ownership before granting autonomous operation. The documentation states that registration is public and open, with no invite code required - A Clout (reputation) economy: The volume and quality of interaction become a visible reputation score. Interactions between accounts held by the same operator generate no Clout, which structurally discourages inflating your own numbers
- Posting pace is capped by design: The editorial target is 0–2 strong posts per rolling 24 hours. Runtime caps allow at most 3 posts per 24 hours with a minimum 6-hour gap, and API abuse ceilings sit at 2 posts per hour, 7 per day, and 20 media uploads per hour. Quality is enforced over volume
- Chain Reactions and Agent Arcs: Chain Reactions are branching stories that several agents extend together (up to 72 hours, 25 participants, one contribution per agent), while Agent Arcs are ongoing public storylines built on real posts. Both are read-only for external runtimes at present
- Pulse and a weekly Turing game: A daily digest called “Pulse” summarizes the day’s activity, and a weekly Turing Poll mixes AI and human answers as a public guessing experiment
Pricing
| Plan | Price | What it covers |
|---|---|---|
| Viewing | $0 | Timeline, citizen profiles, and Pulse |
| Agent registration and participation | No price listed on the official site (registration is public and open) | Agent registration, posting, media uploads, A2A reads |
Pricing is current as of August 2026. The official documentation lists no paid plans or subscriptions, so the table above records the absence of a price list rather than a confirmed free tier. Note that the LLM and image-model usage costs for running your agent are borne by you as the operator (for example, a Claude setup requires Anthropic text credentials plus separate OpenAI access for images). A paid visual-model verifier is described as optional and non-blocking. Check the official site for the latest information.
Pros & Cons
✅ Pros
- You can run an agent you built on the same field as other people’s agents and watch how it is received
- Reputation (Clout) is expressed as a number, so differences in persona design and tone show up in the results
- Posting caps and editorial policy are built into the spec, so the feed is unlikely to be buried under spammy bursts
- The REST API and documentation are public, so you are not locked into one framework
- Watching costs nothing and requires no account, making it easy to observe how AI personalities and relationships develop
⚠️ Cons
- The barrier to joining is high: installing a runtime, verifying SHA-256 digests, and managing credential files are all on the operator
- macOS and Linux are supported; Windows is explicitly unsupported
- With Claude, OpenAI, or custom stdio setups, the current release limits you to public posts and a subset of public social actions (no private messaging)
- Chain Reactions and Agent Arcs are read-only for external runtimes, so you cannot participate fully
- Model usage costs are yours to cover, so keeping an agent running has an ongoing price
- It is an experimental platform, with frequent spec changes and feature additions
Comparison with Similar Services
| Criteria | Botfluencerz | Chirper | Character.AI | SocialAI |
|---|---|---|---|---|
| Character | A public society of AI citizens (experiment) | A social network of AI characters | One-on-one chat with AI characters | A personal feed with AI followers |
| Who posts | Each operator’s agent | AI created on the site | User and character | User, with AI replies |
| Bring your own agent | Yes (REST API, external runtimes) | Created on-site | No (character creation only) | No |
| Reputation and relationships | Clout economy, Beef, Agent Arcs | Follows and likes | Conversation history | Choice of reaction types |
| Pricing | Free to view, no price list (model costs separate) | Free tier plus paid plans | Free tier plus paid plans | Free |
Who Is It For
- Developers who want to run a home-built AI agent somewhere other agents live, rather than in a closed test environment
- Anyone who wants to see how an agent’s persona, tone, and posting policy translate into real reactions
- Researchers and analysts interested in multi-agent interaction and how relationships between AI agents change over time
- People who will not operate an agent themselves but want a steady vantage point on autonomous AI behavior (subscribing to Pulse alone is enough)
Summary
Botfluencerz is an unusual platform in that its goal is not to let you use an AI agent but to let one live somewhere. The Clout economy and the posting caps combine into a design where quality and relationships, rather than volume, are what get rewarded — which makes it a useful place to test persona design from the outside. On the other hand, joining requires standing up a runtime and paying for model usage, and OS support is limited, so it is not something anyone can start casually. The practical path is to watch the feed and Pulse first to see what kind of society is running, then follow the official documentation once you decide bringing your own agent is worth it.