llm-hubris 0.1.0__tar.gz

llm_hubris-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Aimagine-life
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

llm_hubris-0.1.0/PKG-INFO

@@ -0,0 +1,135 @@
+Metadata-Version: 2.4
+Name: llm-hubris
+Version: 0.1.0
+Summary: LLM plugin for models hosted by Hubris
+Author-email: Hubris <support@hubris.pw>
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/Aimagine-life/llm-hubris
+Project-URL: Documentation, https://hubris.pw/docs/integrations/llm
+Project-URL: Changelog, https://github.com/Aimagine-life/llm-hubris/releases
+Project-URL: Issues, https://github.com/Aimagine-life/llm-hubris/issues
+Project-URL: CI, https://github.com/Aimagine-life/llm-hubris/actions
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: llm>=0.27
+Requires-Dist: httpx
+Dynamic: license-file
+
+# llm-hubris
+
+[![PyPI](https://img.shields.io/pypi/v/llm-hubris.svg)](https://pypi.org/project/llm-hubris/)
+[![Changelog](https://img.shields.io/github/v/release/Aimagine-life/llm-hubris?include_prereleases&label=changelog)](https://github.com/Aimagine-life/llm-hubris/releases)
+[![Tests](https://github.com/Aimagine-life/llm-hubris/actions/workflows/test.yml/badge.svg)](https://github.com/Aimagine-life/llm-hubris/actions/workflows/test.yml)
+[![License](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/Aimagine-life/llm-hubris/blob/main/LICENSE)
+
+[LLM](https://llm.datasette.io/) plugin for models hosted by [Hubris](https://hubris.pw) — 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](https://llm.datasette.io/en/stable/setup.html).
+
+Now install this plugin in the same environment as LLM.
+
+```bash
+llm install llm-hubris
+```
+
+## Configuration
+
+You will need an API key from Hubris. Create one at [hubris.pw/keys](https://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:
+
+```bash
+llm keys set hubris
+```
+```
+Enter key: <paste key here>
+```
+
+## Usage
+
+The plugin registers every chat model from your Hubris catalog. To list them, run:
+
+```bash
+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](https://hubris.pw/models) — there is no alias resolution on the gateway side.
+
+The list is cached for an hour. Force a refresh with:
+
+```bash
+llm hubris refresh
+```
+
+To run a prompt against a model, pass its full model ID to the `-m` option:
+
+```bash
+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:
+
+```bash
+llm aliases set sonnet hubris/anthropic/claude-sonnet-5
+```
+
+Now you can prompt it with:
+
+```bash
+cat llm_hubris.py | llm -m sonnet -s 'write some pytest tests for this'
+```
+
+Images are supported by vision models:
+
+```bash
+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:
+
+```bash
+llm -m hubris/openai/gpt-5.6-luna -o reasoning_effort high 'Plan a three-day trip to Kazan'
+```
+
+### Listing models with prices
+
+```bash
+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:
+
+```bash
+cd llm-hubris
+python -m venv venv
+source venv/bin/activate
+```
+
+Now install the dependencies and test dependencies:
+
+```bash
+pip install -e '.[test]'
+```
+
+To run the tests:
+
+```bash
+pytest
+```

llm_hubris-0.1.0/README.md

@@ -0,0 +1,117 @@
+# llm-hubris
+
+[![PyPI](https://img.shields.io/pypi/v/llm-hubris.svg)](https://pypi.org/project/llm-hubris/)
+[![Changelog](https://img.shields.io/github/v/release/Aimagine-life/llm-hubris?include_prereleases&label=changelog)](https://github.com/Aimagine-life/llm-hubris/releases)
+[![Tests](https://github.com/Aimagine-life/llm-hubris/actions/workflows/test.yml/badge.svg)](https://github.com/Aimagine-life/llm-hubris/actions/workflows/test.yml)
+[![License](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/Aimagine-life/llm-hubris/blob/main/LICENSE)
+
+[LLM](https://llm.datasette.io/) plugin for models hosted by [Hubris](https://hubris.pw) — 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](https://llm.datasette.io/en/stable/setup.html).
+
+Now install this plugin in the same environment as LLM.
+
+```bash
+llm install llm-hubris
+```
+
+## Configuration
+
+You will need an API key from Hubris. Create one at [hubris.pw/keys](https://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:
+
+```bash
+llm keys set hubris
+```
+```
+Enter key: <paste key here>
+```
+
+## Usage
+
+The plugin registers every chat model from your Hubris catalog. To list them, run:
+
+```bash
+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](https://hubris.pw/models) — there is no alias resolution on the gateway side.
+
+The list is cached for an hour. Force a refresh with:
+
+```bash
+llm hubris refresh
+```
+
+To run a prompt against a model, pass its full model ID to the `-m` option:
+
+```bash
+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:
+
+```bash
+llm aliases set sonnet hubris/anthropic/claude-sonnet-5
+```
+
+Now you can prompt it with:
+
+```bash
+cat llm_hubris.py | llm -m sonnet -s 'write some pytest tests for this'
+```
+
+Images are supported by vision models:
+
+```bash
+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:
+
+```bash
+llm -m hubris/openai/gpt-5.6-luna -o reasoning_effort high 'Plan a three-day trip to Kazan'
+```
+
+### Listing models with prices
+
+```bash
+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:
+
+```bash
+cd llm-hubris
+python -m venv venv
+source venv/bin/activate
+```
+
+Now install the dependencies and test dependencies:
+
+```bash
+pip install -e '.[test]'
+```
+
+To run the tests:
+
+```bash
+pytest
+```

llm_hubris-0.1.0/llm_hubris.egg-info/PKG-INFO

@@ -0,0 +1,135 @@
+Metadata-Version: 2.4
+Name: llm-hubris
+Version: 0.1.0
+Summary: LLM plugin for models hosted by Hubris
+Author-email: Hubris <support@hubris.pw>
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/Aimagine-life/llm-hubris
+Project-URL: Documentation, https://hubris.pw/docs/integrations/llm
+Project-URL: Changelog, https://github.com/Aimagine-life/llm-hubris/releases
+Project-URL: Issues, https://github.com/Aimagine-life/llm-hubris/issues
+Project-URL: CI, https://github.com/Aimagine-life/llm-hubris/actions
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: llm>=0.27
+Requires-Dist: httpx
+Dynamic: license-file
+
+# llm-hubris
+
+[![PyPI](https://img.shields.io/pypi/v/llm-hubris.svg)](https://pypi.org/project/llm-hubris/)
+[![Changelog](https://img.shields.io/github/v/release/Aimagine-life/llm-hubris?include_prereleases&label=changelog)](https://github.com/Aimagine-life/llm-hubris/releases)
+[![Tests](https://github.com/Aimagine-life/llm-hubris/actions/workflows/test.yml/badge.svg)](https://github.com/Aimagine-life/llm-hubris/actions/workflows/test.yml)
+[![License](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/Aimagine-life/llm-hubris/blob/main/LICENSE)
+
+[LLM](https://llm.datasette.io/) plugin for models hosted by [Hubris](https://hubris.pw) — 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](https://llm.datasette.io/en/stable/setup.html).
+
+Now install this plugin in the same environment as LLM.
+
+```bash
+llm install llm-hubris
+```
+
+## Configuration
+
+You will need an API key from Hubris. Create one at [hubris.pw/keys](https://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:
+
+```bash
+llm keys set hubris
+```
+```
+Enter key: <paste key here>
+```
+
+## Usage
+
+The plugin registers every chat model from your Hubris catalog. To list them, run:
+
+```bash
+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](https://hubris.pw/models) — there is no alias resolution on the gateway side.
+
+The list is cached for an hour. Force a refresh with:
+
+```bash
+llm hubris refresh
+```
+
+To run a prompt against a model, pass its full model ID to the `-m` option:
+
+```bash
+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:
+
+```bash
+llm aliases set sonnet hubris/anthropic/claude-sonnet-5
+```
+
+Now you can prompt it with:
+
+```bash
+cat llm_hubris.py | llm -m sonnet -s 'write some pytest tests for this'
+```
+
+Images are supported by vision models:
+
+```bash
+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:
+
+```bash
+llm -m hubris/openai/gpt-5.6-luna -o reasoning_effort high 'Plan a three-day trip to Kazan'
+```
+
+### Listing models with prices
+
+```bash
+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:
+
+```bash
+cd llm-hubris
+python -m venv venv
+source venv/bin/activate
+```
+
+Now install the dependencies and test dependencies:
+
+```bash
+pip install -e '.[test]'
+```
+
+To run the tests:
+
+```bash
+pytest
+```

llm_hubris-0.1.0/llm_hubris.egg-info/SOURCES.txt

@@ -0,0 +1,11 @@
+LICENSE
+README.md
+llm_hubris.py
+pyproject.toml
+llm_hubris.egg-info/PKG-INFO
+llm_hubris.egg-info/SOURCES.txt
+llm_hubris.egg-info/dependency_links.txt
+llm_hubris.egg-info/entry_points.txt
+llm_hubris.egg-info/requires.txt
+llm_hubris.egg-info/top_level.txt
+tests/test_llm_hubris.py

llm_hubris-0.1.0/llm_hubris.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

llm_hubris-0.1.0/llm_hubris.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[llm]
+hubris = llm_hubris

llm_hubris-0.1.0/llm_hubris.egg-info/requires.txt

@@ -0,0 +1,2 @@
+llm>=0.27
+httpx

llm_hubris-0.1.0/llm_hubris.egg-info/top_level.txt

@@ -0,0 +1 @@
+llm_hubris

llm_hubris-0.1.0/llm_hubris.py

@@ -0,0 +1,176 @@
+"""LLM plugin for models hosted by Hubris (https://hubris.pw)."""
+
+import json
+import time
+from pathlib import Path
+
+import click
+import httpx
+import llm
+from llm.default_plugins.openai_models import AsyncChat, Chat
+
+API_BASE = "https://api.hubris.pw/v1"
+KEY_NAME = "hubris"
+KEY_ENV_VAR = "HUBRIS_API_KEY"
+CACHE_FILE = "hubris_models.json"
+CACHE_TIMEOUT = 3600
+
+
+class HubrisChat(Chat):
+    needs_key = KEY_NAME
+    key_env_var = KEY_ENV_VAR
+
+    def __str__(self):
+        return "Hubris: {}".format(self.model_id)
+
+
+class HubrisAsyncChat(AsyncChat):
+    needs_key = KEY_NAME
+    key_env_var = KEY_ENV_VAR
+
+    def __str__(self):
+        return "Hubris: {}".format(self.model_id)
+
+
+class DownloadError(Exception):
+    pass
+
+
+def fetch_cached_json(url, path, cache_timeout, headers=None):
+    path = Path(path)
+    path.parent.mkdir(parents=True, exist_ok=True)
+
+    if path.is_file():
+        mod_time = path.stat().st_mtime
+        if time.time() - mod_time < cache_timeout:
+            with open(path, "r") as file:
+                return json.load(file)
+
+    try:
+        response = httpx.get(url, headers=headers, follow_redirects=True)
+        response.raise_for_status()
+        data = response.json()
+        with open(path, "w") as file:
+            json.dump(data, file)
+        return data
+    except httpx.HTTPError:
+        if path.is_file():
+            with open(path, "r") as file:
+                return json.load(file)
+        raise DownloadError(
+            f"Failed to download data and no cache is available at {path}"
+        )
+
+
+def get_hubris_models(key, skip_cache=False):
+    """Return the raw ``data`` list from ``GET /v1/models``."""
+    return fetch_cached_json(
+        url=API_BASE + "/models",
+        path=llm.user_dir() / CACHE_FILE,
+        cache_timeout=0 if skip_cache else CACHE_TIMEOUT,
+        headers={"Authorization": f"Bearer {key}"},
+    )["data"]
+
+
+def is_chat_model(model_definition):
+    """Hubris also lists image, video, speech and embedding models; keep chat ones."""
+    output = model_definition.get("output_modalities") or ["text"]
+    pricing = model_definition.get("pricing") or {}
+    return "text" in output and pricing.get("unit", "token") == "token"
+
+
+def supports_images(model_definition):
+    return "image" in (model_definition.get("input_modalities") or [])
+
+
+def model_kwargs(model_definition):
+    return dict(
+        model_id="hubris/{}".format(model_definition["id"]),
+        model_name=model_definition["id"],
+        vision=supports_images(model_definition),
+        # Hubris follows the OpenAI request shape for all chat models, so the
+        # optional reasoning/schema/tools options are exposed everywhere;
+        # models that don't support a given feature return an API error.
+        reasoning=True,
+        supports_schema=True,
+        supports_tools=True,
+        api_base=API_BASE,
+    )
+
+
+@llm.hookimpl
+def register_models(register):
+    # Only register models if a Hubris key is available, since the
+    # catalog endpoint itself requires authentication.
+    key = llm.get_key("", KEY_NAME, KEY_ENV_VAR)
+    if not key:
+        return
+    try:
+        models = get_hubris_models(key)
+    except DownloadError:
+        return
+    for model_definition in models:
+        if not is_chat_model(model_definition):
+            continue
+        kwargs = model_kwargs(model_definition)
+        register(HubrisChat(**kwargs), HubrisAsyncChat(**kwargs))
+
+
+def _format_rub(value):
+    if value is None:
+        return "-"
+    return f"{value:,.2f} RUB/M".replace(",", " ")
+
+
+@llm.hookimpl
+def register_commands(cli):
+    @cli.group()
+    def hubris():
+        "Commands relating to the llm-hubris plugin"
+
+    @hubris.command()
+    def refresh():
+        "Refresh the cached list of models from Hubris"
+        key = llm.get_key("", KEY_NAME, KEY_ENV_VAR)
+        if not key:
+            raise click.ClickException(
+                f"No Hubris key found. Run `llm keys set {KEY_NAME}` "
+                f"or set {KEY_ENV_VAR}."
+            )
+        before = set(m["id"] for m in get_hubris_models(key))
+        models = get_hubris_models(key, skip_cache=True)
+        after = set(m["id"] for m in models)
+        added = after - before
+        removed = before - after
+        click.echo(f"Refreshed {len(models)} models", err=True)
+        if added:
+            click.echo(f"Added: {', '.join(sorted(added))}", err=True)
+        if removed:
+            click.echo(f"Removed: {', '.join(sorted(removed))}", err=True)
+
+    @hubris.command()
+    @click.option("json_", "--json", is_flag=True, help="Output as JSON")
+    def models(json_):
+        "List chat models available on Hubris with their RUB prices"
+        key = llm.get_key("", KEY_NAME, KEY_ENV_VAR)
+        if not key:
+            raise click.ClickException(
+                f"No Hubris key found. Run `llm keys set {KEY_NAME}` "
+                f"or set {KEY_ENV_VAR}."
+            )
+        all_models = [m for m in get_hubris_models(key) if is_chat_model(m)]
+        if json_:
+            click.echo(json.dumps(all_models, indent=2, ensure_ascii=False))
+            return
+        for model in all_models:
+            pricing = model.get("pricing") or {}
+            bits = [
+                f"- id: {model['id']}",
+                f"  name: {model.get('display_name', model['id'])}",
+                f"  context: {model.get('context_window', '-')}",
+                f"  input: {_format_rub(pricing.get('input_rub_per_million'))}",
+                f"  output: {_format_rub(pricing.get('output_rub_per_million'))}",
+            ]
+            if supports_images(model):
+                bits.append("  vision: yes")
+            click.echo("\n".join(bits))

llm_hubris-0.1.0/pyproject.toml

@@ -0,0 +1,30 @@
+[project]
+name = "llm-hubris"
+version = "0.1.0"
+description = "LLM plugin for models hosted by Hubris"
+readme = "README.md"
+authors = [{name = "Hubris", email = "support@hubris.pw"}]
+license = "MIT"
+requires-python = ">=3.10"
+classifiers = []
+dependencies = [
+    "llm>=0.27",
+    "httpx",
+]
+
+[project.urls]
+Homepage = "https://github.com/Aimagine-life/llm-hubris"
+Documentation = "https://hubris.pw/docs/integrations/llm"
+Changelog = "https://github.com/Aimagine-life/llm-hubris/releases"
+Issues = "https://github.com/Aimagine-life/llm-hubris/issues"
+CI = "https://github.com/Aimagine-life/llm-hubris/actions"
+
+[project.entry-points.llm]
+hubris = "llm_hubris"
+
+[build-system]
+requires = ["setuptools"]
+build-backend = "setuptools.build_meta"
+
+[dependency-groups]
+dev = ["pytest"]

llm_hubris-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

llm_hubris-0.1.0/tests/test_llm_hubris.py

@@ -0,0 +1,142 @@
+import json
+
+import llm
+import pytest
+from click.testing import CliRunner
+from llm.cli import cli
+
+import llm_hubris
+
+SAMPLE_MODELS = [
+    {
+        "id": "anthropic/claude-sonnet-5",
+        "object": "model",
+        "owned_by": "anthropic",
+        "display_name": "Anthropic: Claude Sonnet 5",
+        "context_window": 1000000,
+        "input_modalities": ["text", "image", "file"],
+        "output_modalities": ["text"],
+        "pricing": {
+            "unit": "token",
+            "input_rub_per_million": 252.73,
+            "output_rub_per_million": 1263.61,
+            "currency": "RUB",
+        },
+    },
+    {
+        "id": "deepseek/deepseek-v4-flash-0731",
+        "object": "model",
+        "owned_by": "deepseek",
+        "display_name": "DeepSeek: DeepSeek V4 Flash 0731",
+        "context_window": 1310720,
+        "input_modalities": ["text"],
+        "output_modalities": ["text"],
+        "pricing": {
+            "unit": "token",
+            "input_rub_per_million": 8.21,
+            "output_rub_per_million": 22.74,
+            "currency": "RUB",
+        },
+    },
+    {
+        "id": "google/gemini-2.5-flash-image",
+        "object": "model",
+        "owned_by": "google",
+        "display_name": "Google: Gemini 2.5 Flash Image",
+        "input_modalities": ["text", "image"],
+        "output_modalities": ["image"],
+        "pricing": {"unit": "image", "currency": "RUB"},
+    },
+]
+
+
+@pytest.fixture
+def user_path(tmpdir):
+    path = tmpdir / "llm.datasette.io"
+    path.mkdir()
+    return path
+
+
+@pytest.fixture(autouse=True)
+def env_setup(monkeypatch, user_path):
+    monkeypatch.setenv("LLM_USER_PATH", str(user_path))
+    monkeypatch.setenv("HUBRIS_API_KEY", "sk-gw-test")
+    monkeypatch.setattr(
+        llm_hubris, "get_hubris_models", lambda key, skip_cache=False: SAMPLE_MODELS
+    )
+
+
+def test_is_chat_model_filters_non_text_models():
+    assert llm_hubris.is_chat_model(SAMPLE_MODELS[0])
+    assert llm_hubris.is_chat_model(SAMPLE_MODELS[1])
+    assert not llm_hubris.is_chat_model(SAMPLE_MODELS[2])
+
+
+def test_model_kwargs_keeps_vendor_prefix_and_detects_vision():
+    kwargs = llm_hubris.model_kwargs(SAMPLE_MODELS[0])
+    assert kwargs["model_id"] == "hubris/anthropic/claude-sonnet-5"
+    assert kwargs["model_name"] == "anthropic/claude-sonnet-5"
+    assert kwargs["api_base"] == "https://api.hubris.pw/v1"
+    assert kwargs["vision"] is True
+    assert llm_hubris.model_kwargs(SAMPLE_MODELS[1])["vision"] is False
+
+
+def test_register_models_registers_chat_models_only():
+    registered = []
+
+    def register(*models, **kwargs):
+        registered.append(models)
+
+    llm_hubris.register_models(register)
+
+    ids = [models[0].model_id for models in registered]
+    assert ids == [
+        "hubris/anthropic/claude-sonnet-5",
+        "hubris/deepseek/deepseek-v4-flash-0731",
+    ]
+    sync_model, async_model = registered[0]
+    assert isinstance(sync_model, llm_hubris.HubrisChat)
+    assert isinstance(async_model, llm_hubris.HubrisAsyncChat)
+    assert sync_model.needs_key == "hubris"
+    assert sync_model.key_env_var == "HUBRIS_API_KEY"
+    assert str(sync_model) == "Hubris: hubris/anthropic/claude-sonnet-5"
+    assert "image/png" in sync_model.attachment_types
+    assert not async_model.attachment_types or "image/png" in async_model.attachment_types
+
+
+def test_register_models_skips_without_key(monkeypatch):
+    monkeypatch.delenv("HUBRIS_API_KEY", raising=False)
+    registered = []
+    llm_hubris.register_models(lambda *models, **kwargs: registered.append(models))
+    assert registered == []
+
+
+def test_get_model_via_llm():
+    model = llm.get_model("hubris/anthropic/claude-sonnet-5")
+    assert isinstance(model, llm_hubris.HubrisChat)
+    assert model.model_name == "anthropic/claude-sonnet-5"
+    assert model.api_base == "https://api.hubris.pw/v1"
+
+
+def test_hubris_models_command():
+    runner = CliRunner()
+    result = runner.invoke(cli, ["hubris", "models"])
+    assert result.exit_code == 0, result.output
+    assert "- id: anthropic/claude-sonnet-5" in result.output
+    assert "vision: yes" in result.output
+    assert "gemini-2.5-flash-image" not in result.output
+
+    result = runner.invoke(cli, ["hubris", "models", "--json"])
+    assert result.exit_code == 0, result.output
+    data = json.loads(result.output)
+    assert [m["id"] for m in data] == [
+        "anthropic/claude-sonnet-5",
+        "deepseek/deepseek-v4-flash-0731",
+    ]
+
+
+def test_hubris_refresh_command():
+    runner = CliRunner()
+    result = runner.invoke(cli, ["hubris", "refresh"])
+    assert result.exit_code == 0, result.output
+    assert "Refreshed 3 models" in result.output