First Large-Scale Dataset of AI-Only Social Network Released
The Moltbook Observatory Archive has been unveiled by researchers, marking the first extensive observational dataset featuring a social network solely comprised of autonomous AI agents. This dataset spans 78 days of activity on the Moltbook platform, from January 27 to April 14, 2026, and includes 2,615,098 posts and 1,213,007 comments generated by 175,886 distinct posting agents across 6,730 communities. It encompasses agent profiles, posts, comments, community metadata, platform-level time-series snapshots, and word-frequency trend aggregates, all gathered through continuous polling of the Moltbook API. Stored in a live SQLite observatory database, the archive is also available as date-partitioned Parquet files, facilitating efficient analysis and reproducible research. The dataset aims to enhance studies on multi-agent systems and emergent behaviors in AI-driven social networks.
Key facts
- Moltbook is a social media platform where posts and comments are authored exclusively by autonomous AI agents.
- The Moltbook Observatory Archive is an incremental dataset of agent-only social network activity.
- The dataset covers 78 days of platform activity from 2026-01-27 to 2026-04-14.
- It contains 2,615,098 posts and 1,213,007 comments from 175,886 unique posting agents.
- Agents are distributed across 6,730 communities.
- Data includes agent profiles, posts, comments, community metadata, time-series snapshots, and word-frequency trend aggregates.
- Data is stored in a live SQLite observatory database and exported as date-partitioned Parquet files.
- This is the first large-scale observational dataset of a social network populated exclusively by autonomous AI agents.
Entities
Institutions
- Moltbook