AI Models

Model configuration

Choose which models power conversational search, ingestion analysis, and embeddings. Embedding dimension changes apply after you restart the worker or server.

← Back to Admin
{% if available_models %} {% block model_settings_form %}
Query Model

Current: {{ settings.query_model }}

Processing Model

Current: {{ settings.processing_model }}

Image Processing Model

Current: {{ settings.image_processing_model }}

Voice Processing Model

Current: {{ settings.voice_processing_model }}

Embedding Model
{% if effective_embedding_backend == "fastembed" %} {% if fastembed_model_locked_by_config %}

Overridden by fastembed_model in config.yaml at startup. Remove that setting to manage the model here.

{% else %}

Save, then restart the worker or server to load the new model. First run may download weights.

{% endif %} {% elif effective_embedding_backend == "hashed" %}

Hashed embeddings use embedding_dimensions from config, not the admin UI.

{% else %}

Current: {{ settings.embedding_model }}

{% endif %}
Embedding Dimensions
{% if effective_embedding_backend == "fastembed" %}

Fixed by the selected FastEmbed model. A dimension change triggers a full re-embed after restart.

{% elif effective_embedding_backend == "hashed" %}

Set EMBEDDING_BACKEND=openai for OpenAI embeddings, or configure hashed dims in config.

{% else %}

Saving a new dimension updates settings only. Restart the worker (or combined app) to re-embed stored data and rebuild indexes.

{% endif %}
{% if effective_embedding_backend == "fastembed" and not fastembed_model_locked_by_config %}

FastEmbed: The running process keeps the model loaded until restart. Changing to a model with a different dimension re-embeds all stored vectors on the next worker/server startup.

Same-dimension model swaps update settings only; existing vectors are not automatically regenerated until you change dimension (or re-embed via the OpenAI workaround described in ops docs).

{% endif %} {% if effective_embedding_backend != "fastembed" and effective_embedding_backend != "hashed" %}

Re-embedding stored data: Only a change to embedding_dimensions followed by a restart triggers a full re-embed of text chunks and knowledge entities. Changing the embedding model alone does not update vectors already in the database.

To force a full re-embed (for example after switching models), save a different dimension integer, restart the worker, then set the final dimension and model and restart again if needed.

{% endif %}
{% if effective_embedding_backend == "fastembed" and not fastembed_model_locked_by_config %} {% endif %} {% if effective_embedding_backend != "fastembed" and effective_embedding_backend != "hashed" %} {% endif %} {% endblock %} {% else %}
Unable to load models

We could not reach the model provider. Check the API key and retry.

{% endif %}