JanitorAI proxy setup: the four fields, models and price per message
The exact Proxy URL, the model ids to type, what one 2K-in / 300-out turn costs on each, and the proxy errors by code.
A Janitor AI proxy is a Provider setting that sends each message of a chat to a chat completions URL you choose instead of to Janitor's own model. The reply comes from the model id you type, and you pay that API per token. The Proxy URL for RelaySwitch is https://api.relayswitch.io/v1/chat/completions: the full path, with no extra slash and no repeated segment.
If you came from a proxy that started charging or ended its free tier, only three fields change: the URL above, a sk-relay- key from the console, and a model id from the table below. Keys are prepaid from $10 by card or crypto, with no subscription and no KYC. Every reply bills its tokens from that balance and carries a signed receipt, so the price tables on this page are computed from the live catalog, never typed.
01The proxy configuration
Open any chat whose creator allows proxies. Click the gear or menu icon at the top right, choose API Settings, pick Proxy under Provider, then + Add Configuration (older layouts show the same panel behind the using janitor button above the chat, with + New). Five fields:
Config Name: RelaySwitchModel Name: glm-5.3-flash-uncensoredProxy URL: https://api.relayswitch.io/v1/chat/completionsAPI Key: sk-relay-...Custom Prompt: (leave blank)- Config Name: anything; it is the label in the list of configurations.
- Model Name: the id, pasted exactly. Janitor sends it as the
modelfield, so a display name or a stray space gets a 404. - Proxy URL:
https://api.relayswitch.io/v1/chat/completions. Janitor posts to the address as given, so a URL that stops at/v1fails. - API Key: the key, with no space before or after it.
- Custom Prompt: leave it blank. The character card and your persona already go with every message; a prompt here is billed as input on each turn.
Click Add, then Save Settings, then refresh the page. Open the configuration again and check it reads Active. Skip the Check Key/Model button: Janitor's own quickstart calls it unreliable and says to save and refresh instead, and a Network Error after clicking it means nothing. Janitor relays proxy requests from its own servers, so the first message is the real test.
There is no max tokens or temperature field on the proxy form. Those stay in Janitor's generation settings and travel with each request unchanged.
02Which model to type
Any chat id on the Models page goes in Model Name. The 10 builds the catalog tags for uncensored or roleplay use come first, then the flagships on the same key. Prices are USD per million tokens; the last two columns price one 2K-in / 300-out turn and 100 of them. There is no free tier: a $10 top-up covers about 30,179 such turns on GLM-5.3 Flash Uncensored.
| Model | Model Name | Input | Output | one 2K-in / 300-out turn | 100 turns |
|---|---|---|---|---|---|
| GLM-5.3 Flash Uncensored | glm-5.3-flash-uncensored | $0.1086 | $0.3802 | $0.00033 | $0.0331 |
| GLM-5.3 Uncensored | glm-5.3-uncensored | $0.291 | $0.873 | $0.00084 | $0.0844 |
| Gemma 4 31B Fabled | gemma-4-31b-fabled | $0.35 | $1.575 | $0.0012 | $0.1173 |
| Gemma 4 31B Novelist | gemma-4-31b-novelist | $0.35 | $1.575 | $0.0012 | $0.1173 |
| Qwen3.8 27B Novelist | qwen3.8-27b-novelist | $0.35 | $2.10 | $0.0013 | $0.133 |
| Cydonia 24B v4.3 | cydonia-24b-v4.3 | $0.484 | $0.60 | $0.0011 | $0.1148 |
| Uncensored AI Roleplay | uncensored-ai-roleplay | $0.60 | $2.34 | $0.0019 | $0.1902 |
| DeepSeek V4.1 Flash Abliterated | deepseek-v4.1-flash-abliterated | $0.875 | $1.75 | $0.0023 | $0.2275 |
| DeepSeek V4 Pro Uncensored | deepseek-v4-pro-uncensored | $1.00 | $2.00 | $0.0026 | $0.26 |
| MythoMax L2 13B | mythomax-l2-13b | $0.404 | $0.404 | $0.00093 | $0.0929 |
| Claude Sonnet 4.6 | claude-sonnet-4-6 | $1.80 | $9.00 | $0.0063 | $0.63 |
| Claude Opus 5.5 | claude-opus-5-5 | $2.80 | $10.00 | $0.0086 | $0.86 |
| Claude Fable 5 | claude-fable-5 | $3.00 | $15.00 | $0.0105 | $1.05 |
| Gemini 3.1 Pro Preview | gemini-3.1-pro-preview | $1.20 | $7.20 | $0.0046 | $0.456 |
| DeepSeek V3.2 | deepseek-v3.2 | $1.08 | $1.6801 | $0.0027 | $0.2664 |
| Grok 4.6 | grok-4.6 | $0.60 | $1.80 | $0.0017 | $0.174 |
| GLM 5.3 | glm-5.3 | $0.3276 | $1.0296 | $0.00096 | $0.0964 |
| GPT 5 | gpt-5 | $2.50 | $20.00 | $0.011 | $1.10 |
USD, read from the RelaySwitch catalog when this page was built. A turn here is 2,000 tokens in (card, persona and recent history) and 300 out.
A build that thinks before it answers spends part of the reply budget on thinking. If replies come back short or cut off on one, raise max tokens in Janitor's generation settings to 1,000 or more, or pick a build without a thinking step. The uncensored models page lists every tagged build with its daily check, and the roleplay chat runs a character card on the same key in the browser, so you can try a model before typing it into Janitor.
03What a long chat costs
Janitor resends the card, your persona and as much history as fits with every message, so a turn late in a chat sends far more than 2,000 tokens. The same 300-token reply at three history sizes:
| Tokens sent per turn | GLM-5.3 Flash Uncensored | Uncensored AI Roleplay | Claude Sonnet 4.6 |
|---|---|---|---|
| 4,000 | $0.00055 | $0.0031 | $0.0099 |
| 8,000 | $0.00098 | $0.0055 | $0.0171 |
| 16,000 | $0.0019 | $0.0103 | $0.0315 |
USD per turn, live catalog rates. The input side is what grows; the output price only matters for the reply itself.
When a model reports cached input tokens, they bill at its cached rate, which lowers a long chat that resends the same history. The console's Usage page lists every request with its tokens and charge, and each one has a receipt at GET /v1/requests/:id/receipt you can check on the verify page. The pricing page has the billing rules.
04Janitor AI proxy errors by code
Janitor shows the status code the API returned, with its own heading. What each one means here, and the fix:
| Janitor shows | Cause | Fix |
|---|---|---|
| ERROR 400: Bad Request | The model refused a field in the request, or the Custom Prompt made the body invalid. | Clear Custom Prompt, reset generation settings to the defaults, send again. |
| ERROR 401: Invalid Credentials | The key is missing, mistyped, revoked, or has a space around it. | Copy the key again from the console, paste it into API Key, Save Settings, refresh. |
| ERROR 402: Insufficient Credits | The prepaid balance is empty, or the key reached the budget set on it. | Top up, or raise the budget on the key in the console. The next message goes through. |
| ERROR 403 | The key is narrowed to a model list and Model Name is not on it (scope_denied). | Add the model to the key, or change Model Name to one on its list. |
| No endpoints found for model name (404) | model_not_found: the id is misspelled or not in the catalog. route_not_found: the Proxy URL stops at /v1 or has a typo in the path. | Paste the id from the Models page. Set Proxy URL to https://api.relayswitch.io/v1/chat/completions exactly. |
| ERROR 429: Rate Limited | The key hit its requests-per-minute or tokens-per-minute limit. | Wait a few seconds before the next message or swipe, or raise the limit on the key. |
| ERROR 500 / 502 / Provider returned error / Model Down | A host serving the model answered an error. The API hands the request to the next host for that model at the same price; a 503 means none answered. | Send again. If it repeats, check the model on the status page or switch Model Name. |
| Unknown prompt response from worker | Janitor-side: the page lost its saved proxy state. | Refresh the page, confirm the configuration reads Active, send again. |
To separate a key problem from a Janitor setting, list the models the key can call from any terminal:
curl https://api.relayswitch.io/v1/models -H "Authorization: Bearer sk-relay-..."A working key returns a list whose id values are the exact strings for Model Name. The API docs list every error code, and status shows each model's daily canary.
05Sources
Menu path, field names, the full-URL rule and the save-and-refresh flow are from Janitor's help center (help.janitorai.com): the TL;DR Quickstart Proxy Instructions article and its proxy error guides, read 2026-11-02. Error headings are quoted as Janitor prints them. Prices and model ids come from the RelaySwitch catalog at render. The same key runs SillyTavern, Chub AI and RisuAI; only the URL shape differs, and each guide states it.
Questions
- What's a proxy on Janitor AI?
- A Provider option that routes each chat message to a chat completions URL you supply instead of Janitor's own model. The reply comes from the model id in the configuration and the provider bills you per token. Janitor relays the request from its servers, so the provider sees Janitor's address, not yours.
- What proxy should I use for Janitor AI?
- One that publishes its prices per token, gives you your own key with a budget, and runs the models you want. Here one 2K-in / 300-out turn on GLM-5.3 Flash Uncensored costs $0.00033 at today's rate, every reply carries a signed receipt, and a key can be narrowed to a model list with a USD budget so a typo or a long night cannot overspend.
- Is Janitor AI better with a proxy?
- It is different: you choose the model and pay per token instead of using the default. A flagship such as Claude Sonnet writes longer, more consistent prose; an uncensored build answers the card as written; a small roleplay build costs a fraction of either. The table above prices the same turn on each so you can decide by the number.
- How do I allow a proxy on a Janitor AI character?
- The character’s creator decides that when editing the character; the proxy panel cannot override it. If a chat says proxies are not allowed, pick another character, or change the setting on a character you made yourself.
- What API should I use for Janitor AI?
- Any OpenAI-compatible chat completions API works. Janitor needs the full endpoint, so for RelaySwitch the Proxy URL is https://api.relayswitch.io/v1/chat/completions, the key starts with sk-relay-, and Model Name is any chat id on the Models page.
- Is there a free API key for Janitor AI?
- Not here. Keys are prepaid from $10 with no subscription, and that first top-up covers about 30,179 turns of 2,000 tokens in and 300 out on GLM-5.3 Flash Uncensored. Before the first top-up the console chat runs 12 free turns a day, which is enough to try a model before pasting it into Janitor.
- How do I set up my own API in Janitor AI?
- Create a key in the console, open a chat, go to API Settings, choose Proxy, click + Add Configuration, and fill Config Name, Model Name, Proxy URL (https://api.relayswitch.io/v1/chat/completions) and API Key. Click Add, Save Settings, refresh, and confirm the configuration reads Active.
- What do API settings do in Janitor AI?
- They choose where each message goes. Under Provider, the default sends it to Janitor’s own model; Proxy sends it to the URL in your configuration with your key and model id. Generation settings such as max tokens and temperature live in a separate panel and go with the request either way.
- How do I get the proxy URL?
- It is fixed: https://api.relayswitch.io/v1/chat/completions. Nothing is generated per account; only the key is yours. Paste it with no trailing slash and no extra path.
- What is the best AI proxy for Janitor AI?
- Judge by cost per turn at the history size you actually chat at, by whether the model answers your cards without a jailbreak prompt riding along as paid input, and by whether a failed request is retried on another host instead of landing as an error. The long-chat table above shows the first; the uncensored builds answer the second; failover answers the third.
- Is the proxy free to use?
- No. Each reply bills its input and output tokens from a prepaid balance at the rate on the Models page; a failed request is not charged. Top-ups start at $10 by card or crypto.
- Why does Janitor show proxy error 401?
- The API Key field is wrong: a mistyped or revoked key, or a space before or after it. Copy the key again from the console, paste it, Save Settings and refresh.
- Why does Janitor show proxy error 404?
- One of two fields. A Model Name the catalog does not list gets model_not_found; a Proxy URL that stops at /v1 or misspells the path gets route_not_found. Use an id from the Models page and https://api.relayswitch.io/v1/chat/completions exactly.
- Why does Janitor show proxy error 429?
- The key hit its per-minute request or token limit, usually from fast swipes. Wait a few seconds and send again, or raise the limit on the key in the console. A 429 from a host behind the model is retried on another host before you see it.
One Proxy URL, every model
Prepaid from $10 by card or crypto. No KYC, no subscription.