Settings — TrueRef

Settings

Configure TrueRef embedding and indexing options

Embedding Provider

Embeddings enable semantic search. Without them, only keyword search (FTS5) is used.

{#if loading}

Loading current configuration…

{:else}
{#each ['none', 'openai', 'local'] as p} {/each}
{#if provider === 'none'}
Search will use keyword matching only. Results may be less relevant for complex questions.
{/if} {#if provider === 'openai'}
{#each PROVIDER_PRESETS as preset} {/each}
{#if testStatus === 'ok'} Connection successful {#if testDimensions}— {testDimensions} dimensions{/if} {:else if testStatus === 'error'} {testError} {/if}
{/if} {#if provider === 'local'}

Local ONNX model via @xenova/transformers

Model: Xenova/all-MiniLM-L6-v2 · 384 dimensions

{#if localAvailable === null}

Checking availability…

{:else if localAvailable}

@xenova/transformers is installed and ready.

{:else}

@xenova/transformers is not installed. Run npm install @xenova/transformers to enable local embeddings.

{/if}
{/if} {#if saveStatus === 'ok'}
Settings saved successfully.
{:else if saveStatus === 'error'}
{saveError}
{/if}
{/if}

About TrueRef

Self-hosted documentation intelligence platform — a full-stack clone of context7.