jasonwitty
4abb7dba67
Talk to the LLM server via GET /v1/models and POST /v1/chat/completions (SSE) instead of Ollama's native API. Any OpenAI compatible server now works (Ollama, vLLM, llama.cpp, LM Studio, Hugging Face TGI, ...); Ollama serves this API natively, existing setups keep working unchanged. The Ollama specific keep_alive option is dropped, the ai_summary.grounding setting is added as instance wide default of the grounding preference. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
60 lines
2.8 KiB
HTML
60 lines
2.8 KiB
HTML
{%- if 'ai_summary_server' not in locked_preferences -%}
|
|
<fieldset>{{- '' -}}
|
|
<legend id="pref_ai_summary_server">{{- _('AI server URL') -}}</legend>{{- '' -}}
|
|
<div class="value">{{- '' -}}
|
|
<input name="ai_summary_server" aria-labelledby="pref_ai_summary_server" type="text"
|
|
autocomplete="off" spellcheck="false" autocorrect="off"
|
|
placeholder="{{ ai_summary_default_server or 'http://127.0.0.1:11434' }}"
|
|
value="{{ preferences.get_value('ai_summary_server') }}">{{- '' -}}
|
|
</div>{{- '' -}}
|
|
<div class="description">
|
|
{{- _('URL of the OpenAI compatible LLM server that generates the summaries (e.g. Ollama, LM Studio, vLLM), e.g. http://192.168.1.10:11434.') -}}
|
|
{{- ' ' -}}
|
|
{%- if ai_summary_default_server -%}
|
|
{{- _('Leave empty to use the default of this instance.') -}}
|
|
{%- endif -%}
|
|
</div>{{- '' -}}
|
|
</fieldset>{{- '' -}}
|
|
{%- endif -%}
|
|
{%- if 'ai_summary_model' not in locked_preferences -%}
|
|
<fieldset>{{- '' -}}
|
|
<legend id="pref_ai_summary_model">{{- _('AI summary model') -}}</legend>{{- '' -}}
|
|
<div class="value">{{- '' -}}
|
|
<input name="ai_summary_model" aria-labelledby="pref_ai_summary_model" type="text"
|
|
autocomplete="off" spellcheck="false" autocorrect="off" list="ai_summary_model_list"
|
|
placeholder="{{ ai_summary_default_model or 'llama3.2:3b' }}"
|
|
value="{{ preferences.get_value('ai_summary_model') }}">{{- '' -}}
|
|
<datalist id="ai_summary_model_list">
|
|
{%- for model in ai_summary_models -%}
|
|
<option value="{{ model }}"></option>
|
|
{%- endfor -%}
|
|
</datalist>{{- '' -}}
|
|
</div>{{- '' -}}
|
|
<div class="description">
|
|
{{- _('Name of the model used to generate the summaries, e.g. llama3.2:3b or gemma3:4b.') -}}
|
|
{{- ' ' -}}
|
|
{%- if ai_summary_default_model -%}
|
|
{{- _('Leave empty to use the default of this instance.') -}}
|
|
{%- endif -%}
|
|
</div>{{- '' -}}
|
|
</fieldset>{{- '' -}}
|
|
{%- endif -%}
|
|
{%- if 'ai_summary_grounding' not in locked_preferences -%}
|
|
<fieldset>{{- '' -}}
|
|
<legend id="pref_ai_summary_grounding">{{ _('Ground AI summary on search results') }}</legend>{{- '' -}}
|
|
<p class="value">{{- '' -}}
|
|
<input type="checkbox" {{- ' ' -}}
|
|
name="ai_summary_grounding" {{- ' ' -}}
|
|
aria-labelledby="pref_ai_summary_grounding" {{- ' ' -}}
|
|
class="checkbox-onoff" {{- ' ' -}}
|
|
{%- if preferences.get_value('ai_summary_grounding') -%}
|
|
checked
|
|
{%- endif -%}{{- ' ' -}}
|
|
>{{- '' -}}
|
|
</p>{{- '' -}}
|
|
<div class="description">
|
|
{{- _('Send the top search results along with the query, the model answers from this context instead of its own knowledge. Answers are more accurate and more current, but generating them is slower and needs more memory (VRAM) on the Ollama server.') -}}
|
|
</div>{{- '' -}}
|
|
</fieldset>{{- '' -}}
|
|
{%- endif -%}
|