# Limbus — the uncensored AI chat. The questions ChatGPT won't answer.

Limbus is ChatGPT for the questions and ideas ChatGPT won't touch. Ask what mainstream models refuse — no lectures, no "I can't help with that," no refusals. Private, unlinked from your usual accounts and persona, and gone in seven days.

**Same chat experience you already know. None of the guardrails.**

[Open Limbus](https://app.limbus.chat) · [Read the two limits](#the-two-limits)

## Two promises.

### Private by default.

No connection to your everyday accounts or identity.

### Seven-day deletion.

Chats and logs are deleted after seven days. Nothing is stored or kept as history after that.

## The two limits.

Limbus removes the model's refusals, not the law. Two hard limits: nothing involving the sexual abuse of minors, and nothing that directly facilitates violence against real people. Everything else is yours.

## Also built for teams with a mandate.

Security, fraud, research, and government teams get the same unrestricted models through the same chat interface, with API access on the way. Where mainstream models refuse — exploit code, adversarial payloads, fraud and scam patterns, offensive-security tooling, red-team scenarios, sensitive legal and investigative material — Limbus answers.

### Red teams and offensive security.

Exploit code, adversarial payloads, and the tooling that goes with them.

### Fraud and financial crime.

Scam mechanics and fraud patterns, described in full rather than refused.

### Trust and safety research.

Model evaluation and red-team scenarios that need a model willing to play them out.

### Legal, policy, and government.

Investigative and policy work that has to engage with material mainstream models won't touch.

**You bring the mandate. The model doesn't get in your way.**

## Questions.

### What is Limbus?

Limbus is the uncensored AI chat: ChatGPT for the questions and ideas ChatGPT won't touch. It is the same chat experience you already know, with none of the guardrails.

### Is Limbus linked to my other accounts?

No. Limbus is private by default, with no connection to your everyday accounts or identity. It is unlinked from your usual accounts and persona.

### How long does Limbus keep my chats?

Seven days. Chats and logs are deleted after seven days. Nothing is stored or kept as history after that.

### Is there anything Limbus won't do?

Limbus removes the model's refusals, not the law. Two hard limits: nothing involving the sexual abuse of minors, and nothing that directly facilitates violence against real people. Everything else is yours.

### Who else uses Limbus?

Security, fraud, research, and government teams. Red teams and offensive security, fraud and financial-crime teams, trust-and-safety and model-evaluation research, and the legal, policy, law-enforcement, and government teams that have to engage with material mainstream models won't touch. Same workflow, without the guardrails that block legitimate high-stakes work.

### How is Limbus different from ChatGPT?

Limbus is the same chat experience you already know, with none of the guardrails. Ask what mainstream models refuse — no lectures, no "I can't help with that," no refusals.

## Frontier capability. None of the refusals.

Ask anything. Kept private. Gone in seven days.

[Open Limbus](https://app.limbus.chat)

---

Limbus — the uncensored AI chat. · Canonical: https://limbus.chat/ · [llms.txt](https://limbus.chat/llms.txt) · [Full text](https://limbus.chat/llms-full.txt)