Proof & Methodology
Evidence you can inspect.
Published evidence, measurement methods, adoption artifacts, and the proof still being built with customers.
Published evidence.
First-party documentation, public SDK artifacts, and named authorship available today.
Security methodology
Request path, tenant isolation, credit safety, session policy, residency, and vulnerability disclosure.Published · /security · /trust · /complianceNamed authors and review
Owned by Suresh, Senior SEO Engineer, and technically reviewed by Rama Surasani, Founder & CEO.Published · /docs · /aboutGitHub SDK repository
Inspectable SDK source, examples, release history, and multi-language client implementations.Published · github.com/nRouterGateway/nrouter-sdknpm package
Published TypeScript and Node.js SDK package metadata, release history, and install instructions.Published · npm install @nrouter_ai/sdkPyPI package
Published Python SDK distribution with inspectable release history and package metadata.Published · pip install nrouter-sdkVideo demonstrations
Official nRouter YouTube channel for product walkthroughs and benchmark demonstrations.Published · youtube.com/@nRouter
Evidence in progress.
Claims that require customer approval, a real incident record, or an independent audit stay labeled.
Customer case studies
Published only after a customer approves attribution, workload, baseline, and measured results.In progress · /customersIndependent routing benchmark
Publishing our measurement protocol for fixed models, region, concurrency, latency, and errors.In progress · /leaderboard · /blogCost-saving benchmark
Audited savings require a customer baseline; use the cost estimator for scenario planning today.In progress · /cost-calculator · /pricingProvider outage postmortem
Published when a real incident occurs, with timeline, detection, impact, and corrective actions.In progress · /statusIndependent community reviews
First-party social accounts are not labeled as reviews; verified references are added as earned.In progress · /community · Product Hunt
Measurement rules.
Every published performance, cost, or security comparison records the conditions that make it reproducible.
Routing.
Model IDs, provider route, fallback order, region, and routing policy.
Performance.
Warm-up count, concurrency, prompt and response size, latency percentiles, errors, and retries.
Cost.
Provider list price, token counts, platform fees, cache state, and invoice period.
Security.
Threat model, test corpus, false positives, redaction results, retention controls, and reviewer.
Developer tutorials.
Runnable guides across frameworks and tools, with explicit compatibility limits.
LangChain
Python and JavaScript chains, RAG, routing, retries, and cost metadata./docs/frameworks/langchainCrewAI
Multi-agent crews with fallback models, guardrails, and budget isolation./docs/frameworks/crewaiVercel AI SDK
TypeScript and Next.js integration over an OpenAI-compatible endpoint./docs/frameworks/vercel-aiClaude Code
Route supported Claude Code traffic through a compatible gateway endpoint./docs/frameworks/claude-codeCursor
Use supported chat models with a provider key and verify Cursor feature limits./docs/frameworks/cursorOpenClaw
Configure an OpenAI-compatible provider and validate model and tool behavior./docs/frameworks/openclaw
Evidence program.
Share a workload, benchmark, or security review to reproduce with our team.
Contact the team
Share a benchmark, security review, or workload to reproduce and publish with its methodology.sales@nrouter.aiCustomer program
Join the design-partner program and co-author a verified case study from real production usage./customersTrust and compliance
Inspect framework status, self-serve legal documents, and third-party subprocessors./trust · /compliance
- Evidence policy: no unnamed logos, no universal savings number, and no benchmark without a method. Last reviewed October 5, 2026 · written by Suresh, Senior SEO Engineer · technically reviewed by Rama Surasani, Founder & CEO of CloudAct Inc..