# Musubi

> Musubi is the AI-powered Trust & Safety platform for content moderation and fraud detection — helping platforms detect, label, and act on fake accounts, scams, coordinated abuse, and policy-violating content at scale, while learning from their team's decisions instead of replacing human judgment.

Musubi is built around two pillars: **Observability** (know what's happening on your platform, and why) and **Action** (stop harm and enforce your policies at scale). Teams typically start with content moderation or fraud detection and layer on additional capabilities over time. Every automated decision comes with a rationale, and humans keep control over thresholds and escalations.

Musubi is agent-ready: it exposes an **MCP server** and a **CLI** so AI agents can create and manage policies, label content, manage datasets, and run investigations alongside the humans they work with. Musubi also runs a suite of **free Trust & Safety tools** at https://tools.musubilabs.ai.

## Start here

- [A Guide to Musubi](https://www.musubilabs.ai/blog/a-guide-to-musubi): Plain-language overview of what Musubi is, the core concepts, and how the pieces fit together.
- [Musubi homepage](https://www.musubilabs.ai/): Product overview and positioning — "Building blocks for Safety and Fraud teams shipping AI. Custom models, agents, workflows, policies, and observability, knit into one powerful system that learns from your team."

## Product

- [Content Moderation](https://www.musubilabs.ai/content-moderation): Write moderation rules in plain language, set automation thresholds, and let Musubi enforce them at scale — with a clear rationale on every decision and no black boxes. Covers text, images, audio, and video in 150+ languages.
- [Fraud Detection](https://www.musubilabs.ai/fraud): Score whole accounts on thousands of behavioral signals; your team's decisions retrain the model nightly, catching ban evaders, scam rings, and account takeovers before they post anything violative.
- [Content Atlas](https://www.musubilabs.ai/content-atlas): Unsupervised clustering that surfaces emerging trends and unknown harm patterns — know what's actually on your platform, not just what got flagged.
- [Moderation Console (Musubi Coop)](https://www.musubilabs.ai/moderation-console): A full moderation console — review queues, automation rules, dashboards, child-safety reporting, and DSA appeals, all built in.
- [Platform: Action](https://www.musubilabs.ai/platform/action): Act on fake accounts, scam rings, and coordinated abuse. Use top-tier LLMs to label and moderate any content type at scale.
- [Platform: Observability](https://www.musubilabs.ai/platform/observability): Collaborate with a moderation system that gets smarter with every use — real-time observability and agreement monitoring for AI moderation.

## For agents

- [Free Trust & Safety tools](https://tools.musubilabs.ai): A set of free, standalone tools for Trust & Safety teams. Musubi also exposes an MCP server and CLI so agents can create policies, label content, manage datasets, and run investigations — see the guide above for how they fit into the workflow.

## Resources

- [Blog](https://www.musubilabs.ai/blog): Writing on content moderation, fraud, policy, and applied AI for T&S.
- [Case studies](https://www.musubilabs.ai/case-studies): How teams use Musubi in production.
- [Industries](https://www.musubilabs.ai/industries): How Musubi adapts to dating, social, marketplaces, gaming, AI platforms, and advertising.

## Company

- [About](https://www.musubilabs.ai/about): The Trust & Safety, ML, and product veterans building Musubi.
- [Book a demo](https://www.musubilabs.ai/book-a-demo): Talk with the team.

## Optional

- [Privacy Policy](https://www.musubilabs.ai/privacy-policy)
- [Terms of Service](https://www.musubilabs.ai/terms-of-service)
- [Cookie Policy](https://www.musubilabs.ai/cookie-policy)