---
title: "Jan custom endpoint setup for an OpenAI-compatible API"
description: "Add BroadNet Inference to the Jan desktop app as a custom OpenAI-compatible provider, add model IDs by hand, and switch on tools and vision per model."
url: https://broadrouter.com/guides/jan-custom-endpoint/
lang: en
updated: 2026-11-16
publisher: BroadRouter (a BroadNet site); API by BroadNet Inference, https://inference.broadnet.ai
---

# Use Jan with BroadNet Inference as a custom OpenAI-compatible provider

Updated 2026-11-16

In Jan, open Settings, Model Providers and click Add Provider. Choose OpenAI-compatible, name it BroadNet, set Base URL to https://inference.broadnet.ai/v1 and paste your BroadNet key. Then add chat models by hand with the + button, for example broadnet/glm-5.3, and turn on Tools or vision for each model under its capabilities.

Jan is an open-source desktop chat app that runs local models and connects to cloud APIs. Any service that speaks the OpenAI format can be added as a **custom provider**. BroadNet Inference serves standard chat completions at `https://inference.broadnet.ai/v1`. BroadRouter is BroadNet’s own site; the steps follow [Jan’s Custom Endpoints page](https://docs.jan.ai/desktop/remote-models/custom-endpoint) as of Jan 0.8.8.

## Setup

1.  **Get a key.** Create a pay-as-you-go key in the BroadNet console, or buy an hourly key or a plan.
    
2.  **Open the dialog.** Go to **Settings** > **Model Providers** and click the **Add Provider** button next to the list.
    
3.  **Pick the API format.** Choose **OpenAI-compatible**, the default.
    
4.  **Fill in the details:**
    
    | Field | Value |
    | --- | --- |
    | Provider name | BroadNet |
    | Base URL | `https://inference.broadnet.ai/v1` |
    | API key | your BroadNet key |
    
5.  Click **Create**.
    
6.  **Add models by hand.** In the provider’s **Models** section, click **+** and enter each chat model ID exactly, for example `broadnet/glm-5.3`.
    
7.  **Set capabilities.** Click the pencil next to each model and turn on what it supports (table below). Jan can’t detect this for custom providers.
    
8.  **Test it.** Start a new chat, pick the BroadNet model in the model selector and send a short message.
    

Jan fetches `https://inference.broadnet.ai/v1/models` when you save. BroadNet’s docs describe that endpoint as catalogue metadata for monitors, and when we checked it on 15 November 2026 it listed only the two transcription models. Those can’t chat; remove them if they appear.

## Model IDs and capabilities

| Model | ID | Tools | Images | Context |
| --- | --- | --- | --- | --- |
| GLM 5.3 | `broadnet/glm-5.3` | On | Off | 1,048,576 |
| Qwen3.8 27B | `broadnet/qwen3.8-27b` | On | On | 262,144 |
| DeepSeek V4.1 Flash | `broadnet/deepseek-v4.1-flash` | Off (not listed by BroadNet) | On | 1,048,576 |
| BroadNet Gemma 4 VL 26B | `broadnet/gemma-4-26b-a4b` | Off | On | 32,768 |

`broadnet/gemma-4-26b-a4b` is the only chat model that runs on the $5 starting credit, and it is the model behind BroadNet’s [hourly key](https://broadrouter.com/blog/hourly-llm-api-key-explained/) ($0.50 for one hour, one stream). The others need a top-up or, for DeepSeek V4.1 Flash, an [unlimited plan](https://broadrouter.com/guides/unlimited-llm-api-sessions/).

## Settings worth changing

-   **Output limit.** Three of the models reason before they answer, and reasoning counts toward `max_tokens`. If you set an output limit for a BroadNet model, keep it at 1,000 tokens or more.
-   **Samplers.** For custom providers Jan exposes the full set of sampling parameters, and its docs warn that custom endpoints may silently ignore samplers they don’t support. If a setting such as min-P seems to do nothing, that is why.
-   **Fallback keys.** Jan can hold several keys per provider and retries with the next one on 401, 403 or 429. BroadNet’s [Acceptable Use Policy](https://www.broadnet.ai/aup) forbids circumventing rate or concurrency limits, so use extra keys for failover, not to multiply your limits.

## Troubleshooting

-   **404 on every request.** The Base URL is missing `/v1`. Jan’s docs call this the most common mistake.
-   **401 or 403.** The key is wrong, expired or out of allowance. A passing **Test keys** result doesn’t prove the key works, because BroadNet’s `/v1/models` answered without a key when we checked.
-   **Model not found.** Check the ID, including the `broadnet/` prefix. Paid models need a top-up.
-   **Empty answers.** Reasoning used the output budget. Raise the output limit. [Why answers come back empty](https://broadrouter.com/guides/reasoning-tokens-empty-response/).
-   **429.** A pay-as-you-go key allows 30 requests a minute and 4 in flight; an hourly key runs one stream. See [errors and retries](https://broadrouter.com/guides/errors-and-retries/).

## Terms in this guide

[Pay-as-you-go](https://broadrouter.com/glossary/pay-as-you-go/)

Pay-as-you-go is a pricing model in which you pay only for what you use, metered per unit such as tokens, requests or audio minutes, with no subscription or minimum commitment. Most LLM APIs bill this way, from a prepaid balance or a periodic invoice.

[Concurrency](https://broadrouter.com/glossary/concurrency/)

Concurrency is the number of requests a client has in flight at the same time. LLM APIs often cap it per key or plan, separately from requests per minute, because each open request holds GPU capacity until its response has finished.

[Base URL](https://broadrouter.com/glossary/base-url/)

A base URL is the fixed root address of an API, such as https://api.example.com/v1, that a client puts in front of every endpoint path, such as /chat/completions. In OpenAI-compatible SDKs, changing the base URL, together with the API key and model ID, points the same code at a different provider.

[Prepaid credit](https://broadrouter.com/glossary/prepaid-credit/)

Prepaid credit is money added to an API account before use. Each request deducts its cost from the balance, and when the balance runs out, metered requests stop until you top up. It caps spend at what you have paid in.

[Token](https://broadrouter.com/glossary/token/)

A token is the unit of text a language model reads and writes: a word, part of a word, a punctuation mark or a byte sequence produced by the model's tokenizer. LLM APIs count, limit and usually bill usage in tokens.

[API key](https://broadrouter.com/glossary/api-key/)

An API key is a secret string that identifies and authorizes a client calling an API. LLM APIs usually expect it in an Authorization: Bearer header, and anyone who holds the key can make requests that are billed to the account that owns it.

[Chat Completions API](https://broadrouter.com/glossary/chat-completions-api/)

The Chat Completions API is OpenAI's HTTP endpoint, POST /v1/chat/completions, that takes a model name and a list of role-tagged messages and returns the model's next message. Its request and response shape became the common format that OpenAI-compatible providers and SDKs reuse.

[Rate limit](https://broadrouter.com/glossary/rate-limit/)

A rate limit is a cap an API places on how much a client can use it in a time window, such as requests per minute or tokens per minute. Requests over the limit are refused, usually with HTTP 429 Too Many Requests, until usage falls back under it.

## Questions

### Why does Jan only show transcription models after I add BroadNet?

Jan fetches https://inference.broadnet.ai/v1/models when you save. When we checked on 15 November 2026, BroadNet's public /v1/models listed only its two transcription models, which don't work in chat. Remove them and add the chat model IDs by hand.

### Jan's Test keys button says OK. Is my key valid?

Not necessarily. Jan tests keys by calling /models, and BroadNet's /v1/models answered without a key when we checked. Send a chat message to confirm the key.

### Can I use the BroadNet hourly key in Jan?

Yes. The hourly key runs BroadNet Gemma 4 VL 26B, ID broadnet/gemma-4-26b-a4b, with one stream at a time, which suits a single chat window.

### Does Jan detect whether a BroadNet model supports images or tools?

No. Jan's docs say custom providers aren't capability-detected. Turn on image input for DeepSeek V4.1 Flash, Qwen3.8 27B and Gemma 4 VL 26B, and Tools for GLM 5.3 and Qwen3.8 27B.

## Related guides

-   [Connect Open WebUI to BroadNet Inference](https://broadrouter.com/guides/open-webui-openai-api/)
-   [Add BroadNet Inference to LibreChat as a custom endpoint](https://broadrouter.com/guides/librechat-custom-endpoint/)
-   [Connect SillyTavern to BroadNet Inference (Custom OpenAI-compatible)](https://broadrouter.com/guides/sillytavern-custom-openai-endpoint/)
-   [Why a reasoning model returns an empty answer](https://broadrouter.com/guides/reasoning-tokens-empty-response/)
-   [All tools that work with BroadNet Inference](https://broadrouter.com/integrations/): SDKs, coding agents, chat UIs and automation tools, with exact settings for each

Try it on your real workload. New accounts start with $5 of credit for Gemma and transcription. DeepSeek, GLM and Qwen3.8 need a top-up.

[Get started](https://inference.broadnet.ai/account/?utm_source=broadrouter&utm_medium=referral&utm_campaign=guides-jan-custom-endpoint&utm_content=page-cta) ·

**Sources**

-   [Jan docs, Custom Endpoints (checked 2026-11-15)](https://docs.jan.ai/desktop/remote-models/custom-endpoint)
-   [Jan releases on GitHub, v0.8.8 latest (checked 2026-11-15)](https://github.com/janhq/jan/releases)
-   [BroadNet Acceptable Use Policy (checked 2026-11-15)](https://www.broadnet.ai/aup)
-   [BroadNet Inference API reference (checked 2026-11-15)](https://inference.broadnet.ai/docs?utm_source=broadrouter&utm_medium=referral&utm_campaign=guides-jan-custom-endpoint&utm_content=sources)
