Longinus Research.

Commercial Tiers

Start via Cloud API, or deploy inside your private VPC

Every tier supports drop-in OpenRouter & Cohere /api/v1/rerank compatibility and strict zero-data-retention guarantees.

01 · Managed Cloud & OpenRouter

CoreRank Cloud

$0.03 / 1M tokens
  • Rerank at $0.03 / 1M tokens (~$0.15 / 1k searches — 10× cheaper than Cohere)
  • Agent Decisions (corerank-decide) at $0.06 / 1M tokens
  • 40% cheaper than Voyage Rerank 3 ($0.05 / 1M tokens)
  • Drop-in OpenRouter & Cohere REST API
  • Strict Zero Data Retention (ZDR)
Request API Key
03 · Design Partner Program

4-Week Paid Pilot

$7,500 flat (credited to annual)
  • Custom configuration for your top 3 ranking or evaluation workflows
  • Side-by-side accuracy, latency & flip-rate audit vs. your current stack
  • Turnkey private container integrated into your pipeline
  • Clear contractual success gate before annual commitment
Book a Pilot

Platform Add-Ons

CoreCompile Proxy & CoreKV Long-Context Engine

Available as standalone private containers or bundled into your Enterprise VPC deployment.

CoreCompile · Agent Hot-Path Proxy

Cut agent guardrail & routing spend by 80–95%

  • Drop-in OpenAI-compatible /v1/chat/completions proxy
  • 0.32 ms CPU guardrails (99.4% SafeGuard AUC, 742×) & GPU tool routing (100% BFCL Top-1, 85.5% HaluBench)
  • Automatic low-confidence fallback to your existing upstream LLM
  • Included in Enterprise Container ($2,500/node/mo) or $0.01 / 1k offloaded calls
Request CoreCompile Access
CoreKV · Long-Context Serving Acceleration

4×–8× higher concurrency on your existing GPUs

  • 4×–8× KV cache memory reduction for long-context production models
  • Up to 3.07× faster decode attention vs. standard PyTorch SDPA
  • 2.03× higher max-batch throughput without upgrading GPU hardware
  • Custom enterprise licensing & turnkey serving integration
Request CoreKV Evaluation

How Private VPC Deployment Works

What is a Self-Hosted VPC deployment?

A Virtual Private Cloud (VPC) is your company's own isolated private network inside AWS, Google Cloud, Azure, or an on-premises data center. Instead of sending confidential code, contracts, financial records, or clinical notes over the public internet to third-party AI APIs, you deploy the CoreRank container directly inside your own infrastructure.

Cloud API vs. Private VPC

Choose the deployment boundary that fits your security posture

Dimension Managed Cloud / OpenRouter Self-Hosted VPC
Data Path Encrypted HTTPS · Zero Data Retention 100% stays inside your network
Compliance Standard SaaS / ZDR SOC2 · HIPAA · GDPR · Air-gapped
Latency (40 docs) ~60–110 ms (incl. RTT) 26–57 ms (local LAN)
High-Volume Cost $0.03–$0.06 / 1M tokens <$0.01 / 1k amortized
Why Regulated & High-Volume Teams Choose VPC

Zero external data egress, sub-60 ms local speed

  • Strict Data Residency: Documents, source repositories, queries, and agent tool calls never leave your AWS, GCP, Azure, or on-prem perimeter.
  • No Internet Round-Trip: Serving directly next to your search index or agent orchestrator removes public-network jitter and rate limits.
  • Predictable Flat Licensing: At high search or agent volumes, flat per-node licensing eliminates runaway per-token API bills.
# Run CoreRank privately inside your AWS / GCP / Azure VPC
docker run --gpus all --network internal-vpc -p 8091:8091 \
  -e CORERANK_LICENSE_KEY="crk_vpc_..." \
  registry.longinusresearch.com/corerank-runtime:latest

Try Before You Commit

Free 24-Hour Private Benchmark Audit

Send us an anonymized sample from your hardest RAG, evaluation, or triage queue — within 24 hours we return a side-by-side accuracy, latency, and adversarial flip-rate report against your current reranker.

Who We Serve

Built for high-stakes ranking workloads

  • AI Labs & Evaluation Teams: High-throughput candidate ranking, preference scoring, and fast automated judging.
  • Regulated Enterprise Search & RAG: LegalTech, FinTech, Healthcare, and Cybersecurity platforms operating inside private VPCs.
  • Trust, Safety & Operations Triage: Priority queues resilient against SEO spam, verbosity, and adversarial tone.

Direct Founder Contact

Marcio Diaz · Founder, Longinus Research

For enterprise pilots, private VPC trials, OpenRouter / Cloud access, or investor inquiries, reach out directly.

Book an Enterprise Pilot

Benchmark CoreRank, CoreCompile, or CoreKV on your own workload

Submit the form below and our engineering team will set up a 1-week shadow evaluation or VPC container trial.

Direct Pilot Request

Start a 1-week shadow benchmark on your real workload

Tell us your primary bottleneck and deployment target (Cloud API or Private VPC). Zero data leaves your infrastructure during VPC pilots.

Or email directly: marcio.diaz@gmail.com