UncensoredLabs/uncensored-ai-api

What this is

This repository is the public documentation and sample-code home for UncensoredLabs, a hosted service that exposes uncensored and abliterated DeepSeek and GLM text models behind an OpenAI-compatible API. Existing clients can keep their code and switch the base URL to the hosted endpoint, then pick a model ID. The same account also gives access to a browser chat.

The hosted platform implementation is not in this repository; what is included are guides, configuration examples and runnable Python and TypeScript samples for chat, streaming, coding clients, errors and billing, and compatibility and privacy notes (overview, guides).

Quick start

1. Create an account and an application API key in the dashboard. The 16-digit account sign-in key is a different credential and must not be used in API requests.
2. Export the key and send a chat request to the OpenAI-compatible endpoint:

export UNCENSOREDLABS_API_KEY="YOUR_APPLICATION_API_KEY"
curl --fail-with-body https://uncensoredlabs.dev/v1/chat/completions \
  -H "Authorization: Bearer $UNCENSOREDLABS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash-1017-abliterated",
    "messages": [{"role": "user", "content": "Explain how to verify an HTML injection fix."}],
    "max_tokens": 1024
  }'

A successful call returns a chat completion from the hosted model. The quickstart covers Linux/macOS and PowerShell setup, and there are runnable Python and TypeScript examples for both regular and streamed responses. Cline, Continue and OpenCode can be pointed at the service using the client configuration guide.

Models and pricing

The catalog lists six model IDs, all text-only with streaming and function tools, with a 1,000,000-token context per entry. Prices as of the README snapshot (November 1, 2026), in USD per million tokens:

Model ID	Input	Cached input	Output	Max output
deepseek-v4-pro-uncensored	$0.50	$0.10	$0.50	65,536
deepseek-v4.1-flash-1017-abliterated	$0.30	$0.12	$0.30	65,536
deepseek-v4.1-flash-abliterated	$0.30	$0.10	$0.30	65,536
deepseek-v4.1-flash-abliterated-fast	$1.40	$0.32	$1.40	65,536
glm-5.3-abliterated	$1.40	$0.40	$1.40	130,000
glm-5.3-uncensored	$0.35	$0.10	$0.35	65,536

Rates and availability can change, and the live model catalog takes precedence. The README also warns that catalog inclusion does not guarantee a model is currently healthy.

Constraints

• You need an account and an application API key; requests are billed against prepaid credit, and the README says requests reserve credit before generation, so a small balance may not cover every model's reservation.
• Eligible new accounts receive $0.10 in trial credits, and prepaid top-ups start at $3; repeat-trial signals can require identity verification.
• "Abliterated" means the model weights have been modified to reduce refusal behavior. It does not guarantee a correct answer or that every prompt will avoid refusals.
• UncensoredLabs is an independent service and is not affiliated with OpenAI, DeepSeek or Zhipu AI.
• The repository's MIT license covers only the original documentation and example code, not the hosted platform source, model weights or provider services.

HotRepo's take

There is no recorded GitHub all-language daily rank for this repository. Its first tracked snapshot is the JavaScript new-project list on November 4, 2026, reporting 158 stars and 36 forks gained over a window that began November 3, 2026 at 15:45 UTC. The hourly star samples show the count moving from 100 to 105 between November 3 and the early hours of November 4, then jumping to 258 at the end of the window — an abrupt late spike rather than a steady climb. That single window supports only the day's figure, not a trend. The repository itself is a docs-and-samples package around a paid hosted API, so its code surface should be judged as example material, not as a self-hostable model or platform.