--- title: "JannyAI: how to use it, and the model to run the card on" canonical: https://uncensoredllm.ai/guides/jannyai generated: 2026-11-09T16:02:06.800Z source: https://uncensoredllm.ai/guides/jannyai base_url: https://api.uncensoredllm.ai/v1 --- # JannyAI: download a card, run it on any uncensored model JannyAI shows and downloads Janitor AI bot definitions; it has no chat, no proxy and no API. This guide takes a JannyAI card into SillyTavern, Janitor AI or our roleplay chat on an uncensored model. Steps checked October 27, 2026 · prices as of November 9, 2026 JannyAI is where people go to read and download the Janitor AI character cards that Janitor hides. It is a card directory, not a chat site: no model, no proxy field, no API. This guide covers how to use JannyAI to get a card, then the three settings that run that card on an uncensored model in SillyTavern, in Janitor AI's own proxy setting, or in the chat on this site, with what 100 messages cost read from the live catalog. JannyAI's pages were read on 27 October 2026. Already on Janitor AI with a proxy? The [Janitor AI proxy guide](https://relayswitch.io/guides/janitorai) is the shorter path. ## What JannyAI is [JannyAI](https://jannyai.com/) is a directory of Janitor AI character cards. Its title reads "JannyAI - Former JanitorAI.me", and on 27 October 2026 its home page listed 2,159,082 entries with tag filters, a token count per card, and a "Hide Low Quality Bots" switch. Its [About page](https://jannyai.com/about) says why it exists: Janitor AI stopped showing bot definitions, so JannyAI shows "the full bot definition/bot description" and lets you "download it to use in other sites or SillyTavern". Each character page has a Character Definition section and a Download button. That is all it does. JannyAI has no chat window, no model, no proxy setting and no API. Searches for a "janny ai proxy" or a "jannyai api" are looking for Janitor AI's proxy field and for an API to put in it. This guide covers both halves: how to use JannyAI to get the card, and where to run it. ## How to use JannyAI 1. Find the card. Browse jannyai.com by tag, or take a Janitor AI character URL and, in the words of its About page, "replace janitorai.com with jannyai.com". 2. Open the character page and read **Character Definition**: the personality, scenario, example dialogue and first message that Janitor AI hides. 3. Click **Download**. The card downloads as a file you can import into SillyTavern or other clients that take Character Card JSON or PNG. 4. Check the token count JannyAI shows under the card name. It is the size of the definition, and it is sent on every message, so it is most of the bill below. JannyAI's own site is the source for its features and any changes; this guide read it on 27 October 2026. ## Where to run the card: three settings Whichever client you pick, the connection is the same three values. The model id goes last; copy it from the table further down. - Base URL: https://api.uncensoredllm.ai/v1 SillyTavern and most clients; ends in /v1 - Proxy URL: https://api.uncensoredllm.ai/v1/chat/completions Janitor AI and other proxy clients - API key: sk-ullm-live-… Created in the console, shown once - Model: cydonia-24b-v4.3 Any id from the models page, typed exactly ### SillyTavern Import the downloaded card from the character panel. Set API to Chat Completion, Chat Completion Source to Custom (OpenAI-compatible), the Custom Endpoint to `https://api.uncensoredllm.ai/v1` and the key, then Connect and pick the model from the list it loads. The [SillyTavern guide](https://relayswitch.io/guides/sillytavern) has every field and the errors. ### Janitor AI Keep chatting on Janitor AI itself but swap the model: in the chat's API settings choose Proxy, set the URL to `https://api.uncensoredllm.ai/v1/chat/completions`, paste the key and the model id. The [Janitor AI proxy guide](https://relayswitch.io/guides/janitorai) walks through those fields. The card stays on Janitor; JannyAI is only needed when you want to read or keep the definition. ### Our roleplay chat No setup: the [roleplay chat](https://uncensoredllm.ai/roleplay) runs the same builds in the browser with a few free turns a day. For a JannyAI card, paste its definition into the system prompt of the [playground](https://uncensoredllm.ai/playground), which exposes that field. ## Which uncensored model for a JannyAI card The builds below carry the roleplay tag in the catalog, answer without a stock refusal layer, and take the OpenAI chat format every client above speaks. Prices are read from the live catalog when the page renders. | Model | Build | Input $/M | Output $/M | Context | Wire formats | | --- | --- | --- | --- | --- | --- | | [Stheno 8B v3.2](https://uncensoredllm.ai/models/stheno-8b-v3.2) stheno-8b-v3.2 | Roleplay | $0.7035 | $0.7035 | 16K | openai | | [Cydonia 24B v4.3](https://uncensoredllm.ai/models/cydonia-24b-v4.3) cydonia-24b-v4.3 | Roleplay | $0.484 | $0.60 | 131K | openai | | [Uncensored AI Roleplay](https://uncensoredllm.ai/models/uncensored-ai-roleplay) uncensored-ai-roleplay | Roleplay | $0.60 | $2.34 | 128K | openai | | [Anubis 70B v1.1](https://uncensoredllm.ai/models/anubis-70b-v1.1) anubis-70b-v1.1 | Roleplay | $1.085 | $1.085 | 32K | openai | Prices per million tokens, read from the live catalog on 2026-11-09. The rest of the roleplay tunes are on the [models page](https://uncensoredllm.ai/models), and the [daily check](https://uncensoredllm.ai/proof) confirms each id serves the model on its label. ## What 100 messages on a JannyAI card cost The worked example: a 3,000-token card (about what JannyAI's counter showed on typical cards on 27 October 2026), 2,000 tokens of chat history on each message, 300-token replies, 100 messages. | Model | 100 messages | Per message | | --- | --- | --- | | [Stheno 8B v3.2](https://uncensoredllm.ai/models/stheno-8b-v3.2) | $0.37 | $0.0037 | | [Cydonia 24B v4.3](https://uncensoredllm.ai/models/cydonia-24b-v4.3) | $0.26 | $0.0026 | | [Uncensored AI Roleplay](https://uncensoredllm.ai/models/uncensored-ai-roleplay) | $0.37 | $0.0037 | | [Anubis 70B v1.1](https://uncensoredllm.ai/models/anubis-70b-v1.1) | $0.58 | $0.0058 | Computed at render from the catalog prices on 2026-11-09: (3,000 + 2,000) × input price + 300 × output price, times 100. The card is 60% of the input on every message. A card trimmed from 3,000 to 1,500 tokens cuts the bill by more than any model change in the table. Top-ups start at $10 and the balance never expires; every rate is on the [pricing page](https://uncensoredllm.ai/pricing). ## Check the key before the first message This request lists the model ids your key can reach. It costs nothing, and a 200 with a list means the key and the URL are both right. ```bash curl https://api.uncensoredllm.ai/v1/models \ -H "Authorization: Bearer sk-ullm-live-…" ``` A 401 is the key: "No API key was sent." means the field is empty, "The API key is not valid." means it is wrong or cut short. A 404 model_not_found is a typo in the model id. A 402 is an empty balance; the request is not charged. The [docs page](https://uncensoredllm.ai/docs) has the same three settings for every other client. ## Questions ### Is JannyAI the same as Janitor AI? No. Janitor AI is the chat site. JannyAI is a separate directory that shows and downloads Janitor AI bot definitions; its own About page says it was made because Janitor AI hid them (read 27 October 2026). ### How do I chat on JannyAI? You cannot; it has no chat. Download the card and run it in SillyTavern, in Janitor AI through its proxy setting, or paste the definition into the playground here. See the playground. ### Is there a JannyAI proxy or JannyAI API? JannyAI has neither. The proxy field is Janitor AI's; the API that goes in it is this one: https://api.uncensoredllm.ai/v1/chat/completions with your key and a model id. See the Janitor AI proxy guide. ### What is a JannyAI alternative? For the definitions, Chub and SillyTavern card sites hold the same format. For running a card on a model you choose, any client that takes an OpenAI-compatible base URL, with this API behind it. See Character AI alternatives. ### Do I need a GPU? No. The model runs on the API; your client only sends the card and the chat. A GPU matters only if you run the model yourself. ### Which model is closest to Janitor AI’s own? Any roleplay-tagged build takes a Janitor card as written. Start with the 24B tune in the table and move up to the 70B one if you want longer, steadier replies. See the model list. Every model above is on the [uncensored model list](https://uncensoredllm.ai/models) with its live price per million tokens. [Per-token pricing](https://uncensoredllm.ai/pricing) sets the whole catalog side by side, and the [family directory](https://uncensoredllm.ai/uncensored) groups the builds by base model. [Get a key](https://app.uncensoredllm.ai/console) Prepaid from $10, no subscription. ## New models and price changes, once a month. New uncensored builds with their prices, price moves, and what the daily model check found.