Models / Uncensored
Uncensored LLM API
6 uncensored and abliterated models on one OpenAI-compatible key, from $0.084 per million tokens. The same key runs 73 more chat models, Claude and GPT included. Prepaid from $10, card or crypto.
Get a key https://api.uplinkswitch.com/v1
Uncensored models · per million tokens, USD
| Model | Type | Input | Cached input | Output | Daily check |
|---|---|---|---|---|---|
| GLM-5.3 Uncensoredglm-5.3-uncensored | Uncensored | $0.18 | $0.051 | $0.18 | Same fingerprint since Oct 12 |
| GLM-5.3 Abliteratedglm-5.3-abliterated | Abliterated | $2.25 | $0.66 | $2.25 | Same fingerprint since Oct 12 |
| DeepSeek V4 Pro Uncensoreddeepseek-v4-pro-uncensored | Uncensored | $0.75 | $0.15 | $0.75 | Same fingerprint since Oct 12 |
| Roleplay Uncensoredroleplay-uncensored | Roleplay | $0.30list $1.00 | input rate | $0.60list $2.00 | Same fingerprint since Oct 12 |
| GLM-5.3 Flash Abliteratedglm-5.3-flash-abliterated | Abliterated | $0.66 | $0.18 | $0.66 | Same fingerprint since Oct 12 |
| GLM-5.3 Flash Uncensoredglm-5.3-flash-uncensored | Uncensored | $0.084 | $0.0084 | $0.084 | Same fingerprint since Oct 12 |
Connect · base URL and key
Python · OpenAI SDK
from openai import OpenAI
client = OpenAI(
base_url="https://api.uplinkswitch.com/v1",
api_key="sk-uplink-live-...",
)
completion = client.chat.completions.create(
model="glm-5.3-uncensored",
messages=[{"role": "user", "content": "Summarize this in one line."}],
)
print(completion.choices[0].message.content)curl
curl https://api.uplinkswitch.com/v1/chat/completions \
-H "authorization: Bearer sk-uplink-live-..." \
-H "content-type: application/json" \
-d '{"model":"glm-5.3-uncensored","messages":[{"role":"user","content":"Hello"}]}'SillyTavern · API Connections
API Chat Completion
Chat Completion Source Custom (OpenAI-compatible)
Custom Endpoint (Base URL) https://api.uplinkswitch.com/v1
Custom API Key sk-uplink-live-...
Enter a Model ID glm-5.3-abliteratedJanitor AI · Proxy
Provider → Proxy → + New
Name UplinkSwitch
Proxy URL https://api.uplinkswitch.com/v1/chat/completions
API key sk-uplink-live-...
Model roleplay-uncensoredSame key · Claude, GPT and 67 more
What abliterated means
Abliteration removes the direction in a model that makes it refuse. It compares the model's activations on prompts it refuses with prompts it answers, finds the one direction that separates them, and projects it out of the weights. Nothing is retrained, so the model keeps what it knows. Maxime Labonne's write-up on Hugging Face walks through the method.
- Abliterated
- The open weights with the refusal direction removed. GLM-5.3 Abliterated, GLM-5.3 Flash Abliterated.
- Uncensored
- Builds of the open weights changed to answer instead of refuse. GLM-5.3 Uncensored, DeepSeek V4 Pro Uncensored, GLM-5.3 Flash Uncensored.
- Roleplay
- An independent model for character chat. Roleplay Uncensored.
Picking a model
- Roleplay Uncensored: Replies start at once, with no thinking step.
- DeepSeek V4 Pro Uncensored: Thinks briefly, then answers.
- GLM-5.3 Abliterated: Thinks before it answers. Give it 1,000 or more max tokens.
Setup mistakes
- SillyTavern and the OpenAI SDK take the base URL ending in
/v1. Janitor AI takes the full URL ending in/chat/completions. - The key goes in the API key field or an
Authorization: Bearerheader. A key in the URL is refused. - Model ids are exact and lowercase:
glm-5.3-abliterated, not GLM-5.3 Abliterated.GET /v1/modelslists every id the key can call.
Questions
- What is an uncensored LLM API?
- An OpenAI-compatible endpoint that serves uncensored and abliterated models. Here it is https://api.uplinkswitch.com/v1: set it as the base URL, add your key, and call any id in the table, such as glm-5.3-uncensored.
- What does abliterated mean?
- The direction in the model that makes it refuse has been found in its activations and removed from its weights, so it answers prompts it used to decline. Nothing is retrained.
- How do I use an abliterated model?
- Call it like any OpenAI chat model: base URL https://api.uplinkswitch.com/v1, your key, and the model id glm-5.3-abliterated. SillyTavern takes the same two values as a Custom (OpenAI-compatible) source.
- Can't find uncensored models on OpenRouter?
- Every id on this page is an uncensored or abliterated build at one listed price. A client that worked with OpenRouter needs two changes: the base URL https://api.uplinkswitch.com/v1 and an UplinkSwitch key.
- Which uncensored model costs the least?
- GLM-5.3 Flash Uncensored (glm-5.3-flash-uncensored): $0.084 per million input tokens and $0.084 per million output tokens.
- What is the Janitor AI proxy URL?
- https://api.uplinkswitch.com/v1/chat/completions. Put your key in API key and a model id such as roleplay-uncensored in Model.
- Does the same key run Claude and GPT?
- Yes. One key and one prepaid balance reach all 79 chat models, Claude, GPT and Gemini included. The Claude models also take the Anthropic format, so Claude Code works with the same key.
- How do I pay?
- Prepaid credit from $10, by card or crypto through our payment processor. No subscription.