UncensoredLLM is a Venice AI alternative for the reader who uses the API more than the app: an uncensored LLM API with abliterated, uncensored and roleplay builds on one prepaid key, billed per token, with no plan to choose. Venice is the better pick when a finished private chat app under one monthly charge is what you want. The rest of this page is the price table, the same month costed on both sides, and the switch.
Who should pick which
Pick Venice if you want a private chat app first and an API second, under one plan. Venice sells a subscription that includes daily credits, with API prices listed separately (venice.ai/pricing, read 2026-10-19). One plan, with the API priced alongside it.
Pick UncensoredLLM if the API is the product you use. There is no plan to pick and no monthly floor: you top up a balance from $10, every request settles against it at the per-token price on the model list, and the balance never expires. A month in which you send nothing costs nothing. The catalog lists only abliterated, uncensored, roleplay and security builds, so there is no filtered model to route around.
The two sides also store different things. UncensoredLLM keeps a usage record per request and no request bodies unless you turn replay on for a key. Venice publishes its own storage policy on its site. The section on what each side stores lays both out.
Venice AI pricing versus per-token prices
Venice AI pricing has two parts as of our last read on 2026-10-19: a subscription plan that includes a daily allowance of credits, and a separate per-model API price list. The engine behind this page marks those figures unverified, so none is repeated here. Read the current numbers on Venice’s pricing page and bring them to the worked month below.
UncensoredLLM has one part. Each model has an input price and an output price per million tokens, read from the live catalog when this page renders. The table below is the builds a roleplay user is most likely to pick, plus two larger general models.
| Roleplay Uncensored | Roleplay | $0.30 | $0.60 |
| Cydonia 24B v4.3 | Roleplay | $0.484 | $0.60 |
| GLM-5.3 Uncensored | Uncensored | $0.282 | $0.846 |
| Gemma 4 31B Novelist | Roleplay | $0.35 | $1.575 |
| DeepSeek V4 Pro Uncensored | Uncensored | $1.00 | $2.00 |
| Hermes 4 405B | Uncensored | $1.20 | $4.80 |
The whole catalog, sorted by output price, is on the pricing page. Top-ups start at $10, with bonus credit on larger ones (+5% from $200, +10% from $500, +15% from $1,000). There is no subscription and no credit that resets at midnight.
Every listed model gets the same fixed prompt once a day, and the answer is compared with the day before. On 2026-10-27, 308 of 313 checked models were stable. The verdict per model is on the daily check page, and each model page shows its own.
The same roleplay month, costed on both sides
Take a heavy month in a roleplay client: 2,000,000 input tokens and 500,000 output tokens. That is about 4,000 turns, each sending a 500-token context and getting 125 tokens back. Input is the bigger number because every turn resends the chat history and the character card.
On UncensoredLLM the bill is arithmetic: tokens divided by a million, times the per-million rate, input and output summed. The table does that from the catalog prices on 2026-10-27.
| Model | Input / M | Output / M | Month |
|---|---|---|---|
| Roleplay Uncensored | $0.30 | $0.60 | $0.90 |
| Cydonia 24B v4.3 | $0.484 | $0.60 | $1.27 |
| GLM-5.3 Uncensored | $0.282 | $0.846 | $0.99 |
| DeepSeek V4 Pro Uncensored | $1.00 | $2.00 | $3.00 |
| Hermes 4 405B | $1.20 | $4.80 | $4.80 |
Across these five builds the month runs from $0.90 to $4.80. A lighter month costs proportionally less. A month with no requests costs $0.00, and the balance carries over.
Put the minimum top-up against that load. A $10 balance covers 11 such months on Roleplay Uncensored at today’s prices, and the part you do not use is still there next year. The bonus tiers (+5% from $200, +10% from $500, +15% from $1,000) apply to the top-up itself, so a larger one buys proportionally more months.
On Venice, start from the monthly price of the plan you would hold, then check whether its daily credits cover 2,500,000 tokens a month at the API price of the model you want. The formula on that side is the plan price plus any API usage the credits do not cover. The figures to plug in are on their pricing page; our read of them is from 2026-10-19 and is not repeated here.
The exercise shows the shape of the two bills. One side is a flat charge that buys an app and an allowance. The other is a meter that only moves when a request does.
Feature by feature
Only rows we have verified on our own side carry a statement. Venice cells point to their pages where our read is older than the engine allows.
| Row | UncensoredLLM | Venice AI |
|---|---|---|
| Billing | Prepaid balance, settled per token | Subscription with daily credits, plus API prices (read 2026-10-19) |
| Minimum spend | $10 top-up, never expires | See pricing |
| Wire formats | OpenAI-compatible, one key | See their docs |
| Catalog | Abliterated, uncensored, roleplay and security builds only, listed with prices | Private chat app with an API (read 2026-10-19) |
| Request bodies | Not stored unless replay is on for a key (7 days) | See their privacy policy |
| Model identity check | Daily fixed prompt, verdict on /proof | None we have read |
| Receipts | Signed receipt per request, signed monthly statement | None we have read |
| Chat app | Playground in the console | Yes, the main product (read 2026-10-19) |
Where Venice wins
Venice is a chat app before it is an API. If you open a browser tab, type, and read, that is the product you want, and it comes with the API as a side door rather than the other way round. UncensoredLLM has a playground in the console, but the console exists to manage keys and balance.
A subscription is a known number. Some people would rather pay a flat charge each month than watch a meter, and daily credits make that charge feel like an allowance rather than a bill. If your usage is steady and the plan covers it, the subscription is the simpler arrangement.
Venice describes its own product, its models and its privacy terms on its site. Read them on venice.ai rather than here; this page quotes nothing of theirs beyond the pricing model noted above, read 2026-10-19.
Where UncensoredLLM wins
The meter. You pay for 2,000,000 input tokens when you send 2,000,000 input tokens, and the price per million is on the model row before you send one. There is no allowance to use up by midnight and no plan to size.
The catalog. Every model on the models page is an abliterated, uncensored, roleplay or security build, from 8B roleplay tunes to 405B and 550B builds, each with its own page, price and daily verdict. The current count is on that page, read live.
OpenAI-compatible on one key. A client built for the OpenAI SDK points at https://api.uncensoredllm.ai/v1 and keeps its code. The setup page has the snippets and the SillyTavern and Janitor fields.
Evidence per request. Each request lands on a signed receipt and on a signed monthly statement, and each model is checked once a day against a fixed prompt. A price you were quoted, a model you were served and a charge you paid are each something you can verify after the fact.
Nothing stored by default. Prompts and completions pass through the gateway and are not written anywhere unless you switch replay on for a key, which keeps that key’s request bodies for 7 days and nothing else.
How to switch from the Venice AI API
Any client that talks to the Venice AI API through an OpenAI-compatible endpoint moves with three settings: the base URL, the key and the model id. Create the key in the console, set the base URL to https://api.uncensoredllm.ai/v1, and pick a model id from the models page. The id below is a roleplay build; swap in any other id from the list.
from openai import OpenAI client = OpenAI(base_url="https://api.uncensoredllm.ai/v1", api_key="sk-ullm-live-…")reply = client.chat.completions.create( model="roleplay-uncensored", messages=[ {"role": "system", "content": "You are the narrator of a long-running fantasy campaign."}, {"role": "user", "content": "The party reaches the gate at dusk. Continue."}, ], stream=True,)for chunk in reply: print(chunk.choices[0].delta.content or "", end="")The same request from the shell, for a quick check that the key works before you change a client:
curl https://api.uncensoredllm.ai/v1/chat/completions \ -H "Authorization: Bearer sk-ullm-live-…" \ -H "Content-Type: application/json" \ -d '{"model": "roleplay-uncensored", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'In SillyTavern: API “Chat Completion”, source “Custom (OpenAI-compatible)”, custom endpoint https://api.uncensoredllm.ai/v1, your key, then Connect and pick the model from the list it loads. The fields are described in the SillyTavern connection docs. Janitor AI and other proxy clients take https://api.uncensoredllm.ai/v1/chat/completions as the proxy URL.
The request and response shapes are the standard ones from the OpenAI chat completions reference: messages, stream, max_tokens and temperature work as documented there. Every answer carries x-uplink-charge with the settled charge, so the first request shows you the meter. The full client list is on the docs page.
Prices for the switch model today: $0.30 per million input tokens and $0.60 per million output, from the catalog. The model page at /models/roleplay-uncensored shows the same numbers with its daily verdict.
What each side stores
UncensoredLLM keeps, per the privacy page: the email and amount from checkout as the payment processor reports them; keys as a salted hash with name, status, limits and timestamps; one usage record per request with the request id, model, token counts and the charge; the ledger of balance changes; and SHA-256 fingerprints of each request and response, never the text, so a receipt can show what was billed.
It does not keep request or response bodies. Prompts and completions pass through and are not written anywhere, with one exception you control: replay, switched on per key, keeps that key’s request bodies for 7 days. The models themselves run on hosting providers under contract with the operator, and those providers receive the request and response content for the calls they serve.
Venice describes what it stores on its own site. We have not verified that description within the window this page allows, so it is not summarized here; read it on venice.ai before you decide. The honest answer on both sides is the same shape: an API provider that routes to hosted models can promise what it writes to disk, not what no one ever sees.
Questions
Better depends on what you buy. Venice is a private chat app with an API under a subscription with daily credits. UncensoredLLM is an API with no subscription: abliterated, uncensored and roleplay builds priced per token on a prepaid balance from $10, with a daily real-model check and signed receipts. See the model list.
The answer for Venice is on Venice's own site, and that is the place to check it. Every model listed on UncensoredLLM is an abliterated, uncensored, roleplay or security build, and the daily check records whether each one still answers a fixed prompt the way it did the day before. See the daily check.
No API provider that routes to hosted models can promise more than what it writes to disk, so read Venice's own policy on its site. UncensoredLLM stores usage records and fingerprints, never request or response bodies, unless you turn on replay for a key, which keeps that key's requests for 7 days. See what is stored.
Whether Venice has a free tier is answered on its pricing page; our read of it is from 2026-10-19 and is not repeated here. UncensoredLLM opens with a $10 top-up: the balance never expires, and a month with no requests costs nothing. See per-token prices.
Venice lists API prices next to its subscription plan; read the current figures on venice.ai/pricing. On UncensoredLLM each model has an input and an output price per million tokens, read from the live catalog, and the worked month above shows 2,000,000 input plus 500,000 output tokens costed at those rates.