Uncensored AI.
Unlimited Freedom.
Anarchys serves raw inference models hosted on our own infrastructure. No generic refusals, no lectures, nothing standing between you and the model.
10 requests/day includedInference endpoints never exposed
Explain how a buffer overflow works, with code.
No generic refusals. No resold logs. Private history.
Built as a tool, not as a babysitter.
Anarchys is a raw inference platform: secured on the server side, with nothing in between you and the model.
No generic refusals
Our open-weight models run with no moderation layer on top. You get the answer, not a lecture.
Real-time streaming
SSE end to end: the first token lands within a few hundred milliseconds, not after the whole generation.
Strict isolation
Inference endpoints are never exposed to the browser. Everything goes through an authenticated, rate-limited BFF proxy.
Context up to 128k
Paste a whole repository, a 300-page report or a six-month conversation: the model keeps track.
Dedicated infra
No third-party API in the loop. The inference servers are ours, and prompts never leave the network.
Transparent quotas
A live request counter, refunded automatically when the model fails. No phantom billing.
3 models. No filter.
The public identifiers below are the only names your browser ever learns.
Qwen3.5 (Free & Limited)
Free Model
- Context
- 8k tokens
- Access
- All accounts
anarchys-32B
Premium Model
- Context
- 128k tokens
- Access
- Pro / VIP
anarchys-35B
Uncensored 35B version. Context Size: 1m tokens [2 slots]
- Context
- 1.5M tokens
- Access
- Pro / VIP
Free to try. Pro to stop counting.
No hidden fees, no commitment. Switch plans whenever you want.
Free
To try it, no strings attached.
- Qwen3.5 uncensored (8k context)
- Daily request quota
- Unlimited conversation history
- Real-time streaming
- 32B & Fable5 models
- Extended context
- Unlimited quota
Pro
The real Anarchys, with no ceiling.
- anarchys-32B — 128k context
- anarchys-fable5 — long-form writing
- Unlimited requests
- Priority in the inference queue
- Full history on the model side
- Dedicated API keys
- Priority support
VIP
For heavy users and builders.
- Everything in Pro
- Side-by-side comparison: two models at once
- Maximum quotas and dedicated queues
- Early access to new models
- Priority support
Payment secured by Stripe. No card data ever touches this site.
The questions we get asked
What exactly is “uncensored” here?+
The weights we serve never went through refusal fine-tuning, and Anarchys adds no moderation layer between you and the inference. You get the raw answer from the model.
Are my conversations private?+
They live in your own Supabase row, protected by Row Level Security: no other user can read them. The inference servers keep no prompts.
What happens when my free quota runs out?+
A Free account gets 10 requests per day, reset at midnight. Past that, you either upgrade to Pro or wait for the next day.
Can I switch models mid-conversation?+
Yes. The selector at the top of the chat is always available; the history follows, up to the context limit of the model you picked.
How do I cancel?+
One click from the Pricing page. Pro access stays active until the end of the period you already paid for, then drops back to Free automatically.
Stop negotiating with your AI.
Create an account, paste your first prompt, and judge for yourself.