Guides
LLM Provider
Under LLM Provider you connect your own language models, in addition to or instead of Aligia's platform models.
What it is for
Section titled “What it is for”Out of the box, Aligia AI runs with platform models. Some organizations want to use their own contract with an AI provider. Others want their data to stay with their own provider. For this you store your own LLM providers here. LLM stands for „Large Language Model“.
Where in the app
Section titled “Where in the app”Aligia AI › Agenten › LLM Provider. The page is titled „LLM-Anbieter“ (LLM providers).
Key terms
Section titled “Key terms”- Kind – „Anthropic“, „OpenAI“ or „OpenAI-kompatibel“ (OpenAI-compatible). For „OpenAI-kompatibel“ you additionally enter a base URL.
- API key – the access key at the provider.
- Model – a specific model of the provider, with model ID and label.
How it works
Section titled “How it works”A provider has a status: „Nicht konfiguriert“ (Not configured), „Wird geprüft…“ (Checking…), „Verbunden“ (Connected), „Fehler“ (Error) or „Modell fehlt“ (Model missing). You store, replace or remove the API key in the separate account interface. It is stored securely and never displayed in Aligia.
For each provider you maintain the models you want to use. Each model has these details:
- context window and maximum output tokens
- whether it supports controllable reasoning
- whether it accepts images
- optionally prices per million tokens
Each model is checked individually and shows its own status.
The models then appear in the model selection of Polaris and of Agents.
Permissions: only owners and admins may create and change LLM providers and models.
Interplay
Section titled “Interplay”Under Einstellungen › Arbeitsbereich (Automation) admins define organization-wide:
- the Polaris default model
- the models for review, chat titles and extraction
- whether platform models are allowed at all
If „Polaris-Plattformmodelle erlauben“ (Allow Polaris platform models) is off, only your own models are available.