---
title: "LoreBary proxy with your own key and an uncensored model"
canonical: https://unrestrictable.ai/guides/lorebary-proxy
generated: 2026-11-16T02:57:38.825Z
source: https://unrestrictable.ai/guides/lorebary-proxy
base_url: https://api.unrestrictable.ai/v1
---

# LoreBary proxy: run an uncensored model behind it

How LoreBary forwards a proxy URL, your key and a model id to Janitor AI, Chub or Wyvern, which features a custom proxy drops, and the direct route when the model is all you want.

Steps checked November 6, 2026 · prices as of November 16, 2026

## Two ways to run this key behind LoreBary

[Sophia's LoreBary](https://lorebary.com) sits between a roleplay front-end and the model. Its Proxy Portal pitch, read 2026-11-02, is to bring any provider key into the chat app you already use, and it ends "Three fields, no lock-in." Janitor AI, Chub or Wyvern stays your chat window. You give it a LoreBary URL instead of a provider URL, and LoreBary works on each message in transit: slash commands typed into the chat, dice and stat plugins, lorebooks, personas, an automatic summary. The model is yours to supply, and the portal says LoreBary charges nothing for the hop.

So the model is never chosen in LoreBary. It is chosen in whatever LoreBary points at. That gives you two routes, and the choice comes down to whether you want the layer:

- **Only the model matters to you.** Skip LoreBary and paste the API into the front-end itself. The same three values work in Janitor AI, RisuAI, Wyvern and Chub; our [Janitor AI](https://relayswitch.io/guides/janitorai), [RisuAI](https://unrestrictable.ai/guides/risuai-custom-api), [Wyvern](https://unrestrictable.ai/guides/wyvern-chat-api) and [Chub AI](https://unrestrictable.ai/guides/chub-ai) walkthroughs show each field. One hop fewer, one bill, and the console lists every request. No front-end at all? The same ids answer in [/chat](https://unrestrictable.ai/chat) with a few free turns a day and no account.
- **You came for the commands and plugins.** Then LoreBary has to carry the key, and the sections below cover what it forwards, which features survive, and the price per message.

- Proxy or base URL: the URL LoreBary gives you Going direct instead, use https://api.unrestrictable.ai/v1 (or https://api.unrestrictable.ai/v1/chat/completions in an app that posts to the address exactly as typed)
- API key: sk-unr-live-… Made in the console and shown one time; LoreBary forwards it unchanged
- Model: uncensored-ai-roleplay Any id on the Models page, spelled exactly as listed

## What you need

- A LoreBary login, plus your usual front-end logged into the same browser. Find the Proxy Portal at [lorebary.com/connect-proxy](https://lorebary.com/connect-proxy).
- An API key from the [console](https://app.unrestrictable.ai/console). Top-ups start at $10, by card, WeChat Pay or crypto; no plan is required, and bigger top-ups earn bonus credit (+5% at $200, +10% at $500, +15% at $1,000, the tiers the console applies today). Keys begin `sk-unr-` and appear once. Since this one will live inside someone else's settings, give it its own key and a monthly cap.
- One model id from the [Models page](https://unrestrictable.ai/models). Three candidates are priced further down.

## Setting up My Custom Proxy

The portal asks how you plan to use LoreBary and shows two cards. The first, **Connect my current app**, is for people who already keep chats, cards and a paid provider key somewhere else. It leads through **Show me the setup** to **Your own key** and a list of providers with a permanent LoreBary route. Pick one and it hands the front-end three values: a **Model** id "including any suffix", the **API key**, and a **Proxy URL** hosted on `api.lorebary.com`.

On 2026-11-02 this API was not among those providers, so you take the second card, **My Custom Proxy**. The guide at lorebary-guide.carrd.co, a community page read the same day that flags itself as outdated, knows the same form under the heading "Cheating On Us?". Field by field:

1. **Endpoint URL**: `https://api.unrestrictable.ai/v1`. The community guide suggests pasting a provider's curl command or endpoint; the base URL ending in `/v1` is that endpoint.
2. **Path to append**: where a choice is offered, the chat completions path `/chat/completions`.
3. **API type**: **OpenAI**. Gemini and Anthropic are the other options; every chat model here speaks the OpenAI format.
4. **Create Proxy**. LoreBary returns a proxy URL of its own. Paste that URL, not `https://api.unrestrictable.ai/v1`, into the proxy field of the front-end. Key and model id go in their usual fields.

For Janitor AI the path is Settings → Provider **Proxy** → **+ New**: LoreBary's URL in the proxy URL box, then your key, then the model id, **Add**, **Save**, and a page reload (the community guide insists on the reload). rpwithai.com's two LoreBary posts (2025-12-15 and 2026-02-01) add that Chub takes the LoreBary URL without `/v1` while Wyvern needs the `/v1` on the end.

You need a LoreBary sign-in to see the form, and the labels above come from the older guide; LoreBary has renamed fields before. Go by what each box wants: a URL, a path, a type.

## Test the key on its own first

Two services in a row means a failure could be either one. Send this request straight to the API, shaped like a front-end would send it. A 200 proves the key, URL and model id; whatever breaks after that is LoreBary's hop. Swap in your key and change nothing else.

```bash
curl -i https://api.unrestrictable.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-unr-live-…" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored-ai-roleplay",
    "max_tokens": 300,
    "messages": [
      {"role": "system", "content": "You are Oren, a lighthouse keeper on a storm night. Stay in character. Reply in two short sentences."},
      {"role": "user", "content": "There is a light out on the water. Is that a ship?"}
    ]
  }'
```

The reply is in the [chat completions shape](https://platform.openai.com/docs/api-reference/chat). Its `usage` block is what you are billed for, and a response header reports the settled charge. Direct or through LoreBary, each request becomes one row on the console's Usage page, so the hop cannot hide what was spent.

## Features that survive a custom proxy

Here is the catch. The community page (lorebary-guide.carrd.co, read 2026-11-02, outdated by its own account) states that commands, plugins and lorebooks only run when the proxy URL is one of LoreBary's permanent provider routes. On a custom proxy LoreBary passes the chat through and does nothing more. What you end up with: your front-end, the model you typed, and a LoreBary account for its card library and forums. The in-chat commands and plugin mechanics are not part of that route.

That is the whole reason the first route exists. If the layer is what you want, ask LoreBary in its Discord for a permanent route to `api.unrestrictable.ai`. If the model is what you want, drop the hop: the direct guides are a two-minute job and lose nothing.

## Model ids that work

Ids from LoreBary's own provider list (`deepseek-chat` and similar) do not exist here. These three do; each goes into the Model field as printed, and the prices are live:

| Model | Build | Input $/M | Output $/M |
| --- | --- | --- | --- |
| [Uncensored AI Roleplay](https://unrestrictable.ai/models/uncensored-ai-roleplay) uncensored-ai-roleplay | Roleplay | $0.60 | $2.34 |
| [GLM-5.3 Flash Uncensored](https://unrestrictable.ai/models/glm-5.3-flash-uncensored) glm-5.3-flash-uncensored | Uncensored | $0.1053 | $0.3685 |
| [DeepSeek V4 Pro Uncensored](https://unrestrictable.ai/models/deepseek-v4-pro-uncensored) deepseek-v4-pro-uncensored | Uncensored | $0.97 | $1.94 |

Prices per million tokens, read from the live catalog on 2026-11-16.

- `uncensored-ai-roleplay`, the [Uncensored AI Roleplay](https://unrestrictable.ai/models/uncensored-ai-roleplay) page: trained for roleplay, and the id we try first whenever a card is new.
- `glm-5.3-flash-uncensored`, the [GLM-5.3 Flash Uncensored](https://unrestrictable.ai/models/glm-5.3-flash-uncensored) page: the budget pick when a session runs long.
- `deepseek-v4-pro-uncensored`, the [DeepSeek V4 Pro Uncensored](https://unrestrictable.ai/models/deepseek-v4-pro-uncensored) page: if a "DeepSeek through LoreBary" thread brought you here. Give it a response length of at least 1,000; it thinks before it answers, and a shorter cap returns nothing.

The [uncensored LLM leaderboard](https://unrestrictable.ai/blog/uncensored-llm-leaderboard) orders every id by price and by the daily check at [/proof](https://unrestrictable.ai/proof); the [Models page](https://unrestrictable.ai/models) lists them all.

## The bill for 100 messages

Because LoreBary takes no cut (the portal's words: "we add no fee"), you pay only for the tokens your front-end sends, at this catalog's rates. The table costs a Janitor-sized turn of 4,000 tokens in and 300 out, then the same turn deep into a long chat with 16,000 tokens in. Per-million prices are the live catalog's on 2026-11-16.

| Model | Per message | Per 100 | Per 100 at 16k context |
| --- | --- | --- | --- |
| `uncensored-ai-roleplay` | $0.0031 | $0.3102 | $1.0302 |
| `glm-5.3-flash-uncensored` | $0.0005 | $0.0532 | $0.1795 |
| `deepseek-v4-pro-uncensored` | $0.0045 | $0.4462 | $1.6102 |

At today's rate the $10 minimum top-up covers roughly 3,223 messages of that size on `uncensored-ai-roleplay`, and the credit stays good for 12 months. Every model and the bonus tiers are on the [pricing page](https://unrestrictable.ai/pricing).

## Who sees the message

Routed through LoreBary, each message passes two operators. LoreBary's privacy policy, which we read 2026-11-02, keeps request content in server memory for one request and no longer. This API keeps what the [privacy page](https://unrestrictable.ai/privacy) lists: one usage row per request (model, request id, token counts, charge), SHA-256 fingerprints of request and response rather than their text, and request bodies only if you switch replay on for that key, kept 7 days. The key is stored as a salted hash and displayed once.

Give a key that lives in a third party's settings a monthly spend limit in the console. Once it is reached the API returns 402 `monthly_limit_reached` rather than charging on. A separate key per app means revoking one leaves the rest working.

## When it fails

| Symptom | Why | What to do |
| --- | --- | --- |
| 401, `invalid_api_key` | Wrong or revoked key, or the key ended up in the URL box. | Copy it from the console again into the key box. |
| 404, `route_not_found` | Endpoint plus appended path does not come out as `/v1/chat/completions`. | Set the endpoint to `https://api.unrestrictable.ai/v1` and the path to `/chat/completions`, then create the proxy again. |
| 404, `model_not_found` | The Model field carries an id from some other provider. | Use an id from the Models page. |
| Commands and plugins are silent | A custom proxy only relays the chat. | Normal on this route. Request a permanent route from LoreBary or connect direct. |
| Janitor keeps using its own model | The proxy settings were not saved, or no reload happened. | Add, Save, then reload the page. |
| Blank reply | A reasoning model ran out of room while thinking. | Set the response length to 1,000 or higher. |

```
HTTP 401, the key is wrong
{"error":{"message":"The API key is not valid.","type":"authentication_error","param":null,"code":"invalid_api_key"}}

HTTP 404, the URL does not reach the chat completions route
{"error":{"message":"No such route.","type":"not_found_error","param":null,"code":"route_not_found"}}

HTTP 404, the Model field names an id that is not on the list
{"error":{"message":"No such model: deepseek-chat.","type":"not_found_error","param":null,"code":"model_not_found"}}

HTTP 402, the balance is empty
{"error":{"message":"The prepaid balance does not cover this request. Top up at … to continue.","type":"billing_error","param":null,"code":"insufficient_balance"}}
```

## Questions

### What does LoreBary do?

It is Sophia’s roleplay platform with a proxy in front: your front-end (Janitor AI, Chub, Wyvern) talks to LoreBary, LoreBary talks to the model API, and on the way it adds commands, plugins, lorebooks and summaries. The model comes from a key you supply.

### Is there a LoreBary fee on top of my key?

No. The Proxy Portal (read 2026-11-02) says LoreBary adds no fee. You pay the API’s per-token price for whatever the front-end sends, from the balance behind your key.

### Which proxy URL goes into Janitor AI with LoreBary?

The one LoreBary shows after Create Proxy, not the API’s address. Skipping LoreBary, the Janitor proxy URL is https://api.unrestrictable.ai/v1/chat/completions. See the Janitor AI proxy setup.

### Will LoreBary commands and plugins run over a custom proxy?

No, going by the community guide read 2026-11-02 at lorebary-guide.carrd.co (self-described as outdated): those features need one of LoreBary’s permanent provider routes. A custom proxy forwards the chat and nothing else.

### Can LoreBary use an uncensored model?

Yes. Create a My Custom Proxy with this API as the endpoint and an id like uncensored-ai-roleplay as the model, or put the API into the front-end directly, which is the shorter path if the model is all you need. See the Models page.

### Is it safe to give LoreBary my key?

LoreBary’s privacy policy (read 2026-11-02) says request content stays in memory for a single request. Handle the key like any key you hand to a third party: a dedicated key, a monthly limit, and revoke it in the console once you stop using the route.

Every model above is on the [uncensored model list](https://unrestrictable.ai/models) with its live price per million tokens. [Per-token pricing](https://unrestrictable.ai/pricing) sets the whole catalog side by side, and the [family directory](https://unrestrictable.ai/uncensored) groups the builds by base model.

[Get a key](https://app.unrestrictable.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.
