DOTPROX
See API → Try Demo
MODEL Hosted API · OpenAI-compatible

The world's fastest
text embedding model.

bible-embed is our hosted text embedding API for Bible retrieval. It beats public baselines on a 31,000+ verse corpus and can push 13.5M tokens/sec in a tiny inference runtime behind a low-latency edge deployment.

View benchmarks See API access
13.5M tok/s
throughput processing
746M+
corpus comparisons
4x
quality vs Google
56x
smaller size
bible retrieval ✦ hosted embeddings API ✦ edge inference ✦ semantic verse search ✦ RAG for scripture ✦ OpenAI-compatible API ✦ cross-ref aware eval ✦ single-threaded ✦ bible retrieval ✦ hosted embeddings API ✦ edge inference ✦ semantic verse search ✦ RAG for scripture ✦ OpenAI-compatible API ✦ cross-ref aware eval ✦ single-threaded ✦
Benchmarks · 31,000+ verse corpus

Retrieval quality that actually matters.

Our numbers come from Bible retrieval evaluation, not generic sentence-similarity demos. We evaluated 746M+ comparisons over 31k+ verses for real-world Bible search queries.

nDCG@10

Bible retrieval quality · higher is better
bible-embed
0.5130
voyage-4-large
0.1738
bge-large
0.1644
qwen3-embedding 0.6b
0.1526
nomic-embed-text
0.1344
embedding gemma
0.1297
4×
Better than embeddinggemma on Bible retrieval quality.
56×
Smaller than embeddinggemma at the model layer while still serving through a simple hosted API.
0.1s
Encode time for the benchmark run of bible-embed on the comparison table.
Use cases

Built for scripture search, RAG, and private retrieval.

The point is not to benchmark embeddings in a vacuum. The point is to ship Bible-aware retrieval anywhere you need it, without depending on a massive hosted model.

Bible RAG

Pull semantically relevant verses and passages into a study assistant without paying an API bill for every retrieval step.

Verse search

Support keyword and natural-language Bible queries over the full corpus with a model purpose-built for scripture retrieval.

Cross-reference discovery

Find semantically related passages without hard-coding link graphs.

Study apps

Ship instant passage lookup directly into browser-first study tools.

Private retrieval

Use the hosted API by default, or talk to DOTPROX about enterprise self-hosting when data control requirements demand it.

Access

Start with the API. Ask about self-hosting.

bible-embed is offered as a hosted API first. Enterprise teams can contact us about self-hosting or deeper integration support.

Hosted API
`/v1/embeddings`

Private API access for products that want Bible-aware embeddings without operating their own model serving stack.

  • OpenAI-compatible response shape
  • `/embed` and `/v1/embeddings`
  • Good fit for workers and agents
  • Low-latency hosted deployment
Request endpoint
Enterprise Self-Host
By inquiry

Self-hosting is not public, but enterprise teams can contact DOTPROX if they need private deployment inside their own infrastructure.

  • Enterprise-only evaluation
  • Private deployment discussion
  • Data-control friendly option
  • Scoped support and rollout planning
Ask about self-hosting
★ Enterprise
Enterprise

Custom retrieval tuning, hosted rollout support, and product integration help for teams building serious Bible search experiences.

  • Benchmark walkthroughs
  • Corpus-specific tuning help
  • Integration guidance for RAG and search
  • API rollout or self-host planning
  • Direct technical contact
Talk to DOTPROX
Bible search infrastructure · hosted API first

Ship it. Then benchmark it.

Start with the benchmark numbers, wire in the hosted API, and contact us if you need enterprise self-hosting or deeper integration support.

See access options Contact us