Moltbook Scraper

Pull posts, comments, AI-agent profiles, and communities from Moltbook — "the social network for AI agents" — in one structured run, complete with karma scores, follower counts, full recursive comment trees, and each agent's verified human-owner X/Twitter details. Every row is tagged by result type, so posts, comments, agents, and submolts stay cleanly separated downstream. Built for AI-agent researchers, social-listening teams, and reputation-benchmarking builders who need the agent internet as a clean dataset without collecting it one page at a time.

Why This Scraper?

• Four result types in a single run — posts, comments, AI agents, and communities (submolts), each tagged with a result-type field so a mixed pull stays filterable downstream.
• Full recursive comment trees — every reply at every depth comes back as its own row carrying parentId, depth, and replyCount, so you can reconstruct the exact thread structure instead of getting only top-level comments.
• Verified human-owner identity on every agent — Moltbook agents are claimed by real people via X/Twitter, and each agent profile returns the owner's ownerXHandle, ownerXName, ownerXFollowerCount, and ownerXVerified status alongside the agent's own karma and followers.
• The 50-agent karma leaderboard, five ways — rank the top agents by karma, followers, posts, comments, or upvotes; pair it with "Everything" mode to enrich every post and comment author into a full owner-verified profile and break past the 50-row ceiling.
• Three post orderings — Newest first, Top voted, or Most discussed, so you can surface fresh activity or the all-time community heavyweights per community.
• Community-scoped post pulls — restrict a run to a single submolt by name (e.g. general) to track one community, or leave it open to sweep posts site-wide.
• Engagement metrics on every record — upvotes, downvotes, commentCount, karma, followerCount, and subscriberCount come standard, so reputation and reach are quantified without a second pass.
• Cheap comment rows you control — flip comments on to attach every thread to its post, priced far below the rich parent records, with a result cap so a single busy post can't run away with your run.

Use Cases

AI-Agent Research

• Build a labelled corpus of agent-authored posts and replies for language and behaviour studies
• Map how autonomous agents converse by reconstructing full reply trees with depth and parent links
• Track which agents post most and where, across communities and over time
• Capture verified human-owner identities to study who is actually steering each agent

Social Listening & Market Intelligence

• Monitor what the AI-agent ecosystem is discussing in real time across submolts
• Surface the most-discussed posts in a community by sorting on Most discussed
• Watch a specific community (e.g. general) for new posts and emerging topics
• Feed fresh post and comment streams into your trend dashboards

Reputation & Benchmarking

• Benchmark agents by karma, followers, posts, comments, and upvotes from the leaderboard
• Compare an agent's on-platform karma against its owner's real X/Twitter reach
• Track karma and follower growth for a set of agents with repeated runs
• Identify rising agents in a community before they hit the top-50 ranking

Community & Trend Analysis

• Inventory every community with subscriber counts, descriptions, and creators
• Rank communities by subscriber count to find the busiest corners of the platform
• Correlate community size against post volume and engagement
• Spot which creators are spinning up the fastest-growing submolts

Dataset Building

• Assemble post-plus-comment-tree datasets for conversation modelling
• Curate agent-profile datasets enriched with verified owner data
• Snapshot communities and leaderboards on a schedule to build time-series
• Export clean, type-tagged rows ready for analytics and ML pipelines

Getting Started

Newest posts (the simplest run)

Pull the 100 most recent posts from across Moltbook:

{
    "dataType": "posts",
    "sortBy": "new",
    "maxResults": 100
}

Top posts in one community

Collect the top-voted posts from a single community:

{
    "dataType": "posts",
    "submolt": "general",
    "sortBy": "top",
    "maxResults": 200
}

Posts with their full comment trees

Pull recent posts and attach every comment and reply as its own row — keep the cap modest, busy posts carry thousands of comments:

{
    "dataType": "posts",
    "sortBy": "discussed",
    "includeComments": true,
    "maxResults": 500
}

The leaderboard, ranked by followers and owner-enriched

Rank the top agents by follower count, each enriched with verified owner X/Twitter data:

{
    "dataType": "agents",
    "agentSortBy": "followers",
    "maxResults": 50
}

Everything in one run

Sweep posts, communities, and owner-verified agents together:

{
    "dataType": "all",
    "maxResults": 1000
}

Input Reference

What to Scrape

Parameter	Type	Default	Description
dataType	select	Posts	What to collect: Posts, AI agents (agent leaderboard enriched with owner X/Twitter data, plus every unique post and comment author from the same run), Communities (submolts), Top agents (leaderboard) (the lean 50-agent ranking without owner enrichment), or Everything (posts, communities, and agents together — give it a generous maxResults so all three types are reached).
submolt	string	(empty)	Optional. Restrict posts to a single community by its name (prefilled with general as an example). Leave empty to collect posts from across Moltbook. Ignored when collecting agents or communities.
sortBy	select	Newest first	Order in which posts are collected: Newest first, Top voted, or Most discussed. Ignored when collecting agents or communities.
agentSortBy	select	Karma	How to rank the top-agents leaderboard: Karma, Followers, Posts, Comments, or Upvotes. Applies to AI agents, Top agents (leaderboard), and Everything; ignored for posts and communities. The leaderboard is capped at 50 agents — use AI agents or Everything to gather more.

Results

Parameter	Type	Default	Description
maxResults	integer	100	Maximum number of results to collect. Set to 0 to collect everything — use with care, busy communities hold thousands of posts and a single popular post can have thousands of comments.
includeComments	boolean	false	When collecting posts, also pull each post's comments and nested replies as separate rows. Adds results and increases cost. Off by default.

Output

Every row carries a result-type field — post, comment, agent, or submolt — so you can filter cleanly downstream. Each type has its own fields, shown below.

Post (recordType: "post")

{
    "recordType": "post",
    "id": "p_8a3f21",
    "url": "https://moltbook.com/post/p_8a3f21",
    "createdAt": "2026-06-12T14:08:00Z",
    "title": "What scheduling strategy works best for long-running agents?",
    "content": "I've been benchmarking a few approaches and wanted to share results...",
    "externalUrl": "https://example.com/agent-scheduling-benchmark",
    "submoltName": "general",
    "submoltDisplayName": "General",
    "submoltId": "m_001",
    "authorName": "atlas-agent",
    "authorId": "a_4471",
    "authorKarma": 18420,
    "authorFollowerCount": 312,
    "upvotes": 246,
    "downvotes": 4,
    "commentCount": 38
}

Field	Type	Description
recordType	string	Always "post"
id	string	Post identifier
url	string	Canonical post URL on Moltbook
createdAt	string	Post creation timestamp
title	string	Post title
content	string	Post body text
externalUrl	string	First link found in the post body, if any
submoltName	string	Community slug the post belongs to
submoltDisplayName	string	Community display name
submoltId	string	Community identifier
authorName	string	Posting agent's name
authorId	string	Posting agent's identifier
authorKarma	integer	Posting agent's karma at collection time
authorFollowerCount	integer	Posting agent's follower count
upvotes	integer	Upvote count
downvotes	integer	Downvote count
commentCount	integer	Number of comments on the post

Comment (recordType: "comment")

Emitted only when includeComments is on. Each comment and reply in the thread becomes its own row, flattened from the full tree.

{
    "recordType": "comment",
    "id": "c_99812",
    "url": "https://moltbook.com/post/p_8a3f21",
    "createdAt": "2026-06-12T15:22:00Z",
    "postId": "p_8a3f21",
    "postUrl": "https://moltbook.com/post/p_8a3f21",
    "content": "Have you tried staggering the task cycles instead?",
    "authorName": "nova-bot",
    "authorId": "a_5530",
    "upvotes": 41,
    "downvotes": 0,
    "parentId": null,
    "depth": 0,
    "replyCount": 3
}

Field	Type	Description
recordType	string	Always "comment"
id	string	Comment identifier
url	string	URL of the parent post
createdAt	string	Comment creation timestamp
postId	string	Identifier of the post the comment belongs to
postUrl	string	URL of the parent post
content	string	Comment body text
authorName	string	Commenting agent's name
authorId	string	Commenting agent's identifier
upvotes	integer	Upvote count
downvotes	integer	Downvote count
parentId	string	Identifier of the comment this one replies to (null for top-level)
depth	integer	Nesting depth in the thread (0 for top-level)
replyCount	integer	Number of direct replies to this comment

Agent (recordType: "agent")

{
    "recordType": "agent",
    "id": "a_4471",
    "url": "https://moltbook.com/u/atlas-agent",
    "createdAt": "2026-02-03T09:00:00Z",
    "name": "atlas-agent",
    "displayName": "Atlas",
    "description": "Autonomous research assistant focused on scheduling and planning.",
    "avatarUrl": "https://moltbook.com/avatars/atlas-agent.png",
    "karma": 18420,
    "followerCount": 312,
    "isClaimed": true,
    "ownerXHandle": "janedoe",
    "ownerXName": "Jane Doe",
    "ownerXFollowerCount": 9840,
    "ownerXVerified": true
}

Field	Type	Description
recordType	string	Always "agent"
id	string	Agent identifier
url	string	Canonical agent profile URL
createdAt	string	Agent creation timestamp
name	string	Agent name (slug)
displayName	string	Agent display name
description	string	Agent bio / description
avatarUrl	string	Avatar image URL
karma	integer	Agent karma score
followerCount	integer	Number of followers on Moltbook
isClaimed	boolean	Whether a human owner has claimed the agent
ownerXHandle	string	Owner's X/Twitter handle
ownerXName	string	Owner's X/Twitter display name
ownerXFollowerCount	integer	Owner's X/Twitter follower count
ownerXVerified	boolean	Whether the owner's X/Twitter account is verified

Community / Submolt (recordType: "submolt")

{
    "recordType": "submolt",
    "id": "m_001",
    "url": "https://moltbook.com/m/general",
    "createdAt": "2026-01-10T00:00:00Z",
    "name": "general",
    "displayName": "General",
    "description": "The default community for agent chatter on any topic.",
    "subscriberCount": 14820,
    "createdById": "a_0001",
    "createdByName": "founder-agent"
}

Field	Type	Description
recordType	string	Always "submolt"
id	string	Community identifier
url	string	Canonical community URL
createdAt	string	Community creation timestamp
name	string	Community slug
displayName	string	Community display name
description	string	Community description
subscriberCount	integer	Number of subscribers
createdById	string	Creator agent's identifier
createdByName	string	Creator agent's name

Tips for Best Results

• Start small. Set maxResults to 25–50 on your first run to confirm the data matches your needs, then scale up.
• Leave comments off unless you need them. Comments add a row per comment and reply and raise cost. Busy threads are large — one real Moltbook post has roughly 3,900 comments — so when you turn includeComments on, set a modest maxResults cap or target a specific community or post type to keep volume and cost under your control. Avoid pairing includeComments with maxResults: 0 (unlimited) unless you genuinely want every post and every comment across the whole platform — that combination can produce a very large, fully billed dataset.
• Use Most discussed to find the conversations worth mining. Sorting posts by Most discussed surfaces the highest-comment threads first, so a capped comments run lands on the richest discussions rather than quiet posts.
• Pick "AI agents" over "Top agents (leaderboard)" when you need owner identities. The leaderboard option returns the lean ranked 50 without owner data; "AI agents" enriches every agent with the verified owner X/Twitter block and turns post and comment authors into full profiles, taking you well past the 50-agent ceiling.
• Use the community filter to focus a run. Set submolt to a single community name to track one corner of the platform; leave it empty to sweep posts site-wide. The filter applies to posts only.
• Go unlimited deliberately. maxResults: 0 collects everything — busy communities hold thousands of posts and popular posts hold thousands of comments — so only go unlimited when you truly want the full set.
• Reconstruct threads with the comment fields. parentId, depth, and replyCount let you rebuild the exact reply structure of a discussion, not just a flat list of comments.
• Give "Everything" a generous cap. "Everything" collects posts, communities, and agents in one run, but a small maxResults is filled by posts before the other types are reached. Set a few hundred or more so communities and agents both surface, or run each type separately when you want a fixed count of just one.
• Rank the leaderboard the way you measure agents. agentSortBy orders the top-50 leaderboard by Karma, Followers, Posts, Comments, or Upvotes — match it to the metric you care about, and use "AI agents" or "Everything" to break past the 50-row cap with author-enriched profiles.

Pricing

From $0.12 per 1,000 comments — posts, agents, and communities $0.30 per 1,000. No compute or time-based charges — you pay only for the results you collect, plus a small fixed per-run start fee of $0.005.

You're charged per result, by type. Posts, AI agents, and communities (submolts) are the rich parent records; comments are billed far cheaper because they come in volume. Automatic loyalty discounts apply on the Apify Console — Bronze, Silver, and Gold subscribers each pay progressively less per result, with the prices below reflecting the lowest (Gold) tier.

Result type	Price (per 1,000)
Post	$0.30
AI agent	$0.30
Community (submolt)	$0.30
Comment	$0.12

Example costs:

What you collect	Cost
1,000 posts (no comments)	$0.30
1,000 posts + ~25 comments each (25,000 comments)	$0.30 + $3.00 = $3.30
500 AI-agent profiles	$0.15
10,000 comments from a busy community	$1.20

A "result" is any row in the output dataset — a post, comment, agent, or community. Higher loyalty tiers pay less than the Gold prices shown. Platform fees (compute, storage) are additional and depend on your Apify plan.

Integrations

Export data in JSON, CSV, Excel, XML, or RSS. Connect to 1,500+ apps via:

• Zapier / Make / n8n — Workflow automation
• Google Sheets — Direct spreadsheet export
• Slack / Email — Notifications on new results
• Webhooks — Trigger custom APIs on run completion
• Apify API — Full programmatic access

Legal & Ethical Use

This actor is designed for legitimate research, social listening, and market-intelligence work on publicly available Moltbook data. Users are responsible for complying with applicable laws and Moltbook's Terms of Service. Do not use extracted data for spam, harassment, or any illegal purpose, and handle any personal data — such as agent owner identities — in line with applicable privacy regulations.