llm-hubris

LLM plugin for models hosted by Hubris — an OpenAI-compatible LLM gateway billed in Russian rubles, with 400+ models (OpenAI, Anthropic, Google, DeepSeek, Qwen and more) behind a single API key.

Installation

First, install the LLM command-line utility.

Now install this plugin in the same environment as LLM.

llm install llm-hubris

Configuration

You will need an API key from Hubris. Create one at hubris.pw/keys.

You can set that as an environment variable called HUBRIS_API_KEY, or add it to the llm set of saved keys using:

llm keys set hubris

Enter key: <paste key here>

Usage

The plugin registers every chat model from your Hubris catalog. To list them, run:

llm models list

You should see a list that looks something like this:

Hubris: hubris/anthropic/claude-sonnet-5
Hubris: hubris/openai/gpt-5.6-luna
Hubris: hubris/google/gemini-3.7-flash
Hubris: hubris/deepseek/deepseek-v4-flash-0731
...

Model IDs always use the full vendor/model form shown in the Hubris catalog — there is no alias resolution on the gateway side.

The list is cached for an hour. Force a refresh with:

llm hubris refresh

To run a prompt against a model, pass its full model ID to the -m option:

llm -m hubris/anthropic/claude-sonnet-5 "Five spooky names for a pet tarantula"

You can set a shorter alias for a model using the llm aliases command:

llm aliases set sonnet hubris/anthropic/claude-sonnet-5

Now you can prompt it with:

cat llm_hubris.py | llm -m sonnet -s 'write some pytest tests for this'

Images are supported by vision models:

llm -m hubris/anthropic/claude-sonnet-5 'describe this image' -a https://static.simonwillison.net/static/2024/pelicans.jpg

Reasoning models accept the reasoning_effort option:

llm -m hubris/openai/gpt-5.6-luna -o reasoning_effort high 'Plan a three-day trip to Kazan'

Listing models with prices

llm hubris models

Outputs every chat model with its context window and RUB prices per million tokens. Add --json for the raw catalog entries.

Development

To set up this plugin locally, first checkout the code. Then create a new virtual environment:

cd llm-hubris
python -m venv venv
source venv/bin/activate

Now install the dependencies and test dependencies:

pip install -e '.[test]'

To run the tests:

pytest