Production-ready retrieval in under 10 minutes. Upload your docs, query in natural language, ship AI features — not infrastructure.
import { RAGClient } from "@ragex/sdk";
const client = new RAGClient({ apiKey: process.env.RAG_API_KEY });
const { results } = await client.search("kb_01", {
query: "What is our refund policy?",
});
// → "Enterprise customers may request a full refund
// within 30 days of purchase..."Trusted by engineering teams
Define a collection to hold your documents. One line of code.
const kb = await client.createKnowledgeBase({
name: "support-docs",
});Upload PDFs, Markdown, HTML — we handle parsing, chunking, and embedding.
const doc = await client.uploadDocument(
kb.id,
pdfFile,
{ filename: "refund-policy.pdf" },
);Ask questions in plain English. Get ranked, relevant results.
const { results } = await client.search(kb.id, {
query: "What is our refund policy?",
});
// → "Enterprise customers may request a full refund
// within 30 days of purchase..."A complete retrieval pipeline so you can focus on your product, not your infrastructure.
PDF, DOCX, PPTX, HTML, Markdown, and 11 more formats — parsed and chunked automatically.
Vector, keyword, or combined search modes with tunable fusion for maximum relevance.
Every query is re-scored with a cross-encoder model to push the best results to the top.
Filter results by custom document metadata using operators like $eq, $in, $gt, and more.
TypeScript and Python SDKs with full type safety and autocompletion.
Isolate knowledge bases per customer. Built for SaaS from day one.
Monitor queries, pages processed, and API requests. Rate limits enforced per plan.
Machine-readable API spec. Generate clients in any language, explore with Swagger.
If reranking times out, results still return. Your search never fails silently.
Clean APIs, typed SDKs, and documentation that respects your time.
“We went from 3 weeks of building RAG infrastructure to shipping in an afternoon. The search quality is genuinely better than what we had in-house.”
Sarah Chen
Staff Engineer, Nimbus
“The API is exactly what I wanted to build but didn't have time to. Clean, fast, and the docs are actually good. Rare combo.”
Marcus Rodriguez
CTO, Vela Labs
“We handle 2M+ queries/month through Ragex. Sub-200ms p95 latency, zero downtime. It just works.”
James Park
Lead Engineer, Stratos
“Replaced our entire Pinecone + LangChain pipeline with a single API call. Our team can focus on the product again.”
Priya Sharma
Engineering Manager, Arcturus
From customer-facing chatbots to internal tooling — one API handles them all.
Ground LLM responses in your actual support docs and knowledge base articles.
Let employees search company wikis, handbooks, and Confluence in natural language.
Semantic search across your docs — better than keyword matching, zero training required.
Search contracts, policies, and regulatory documents with cited sources.
Build in-app AI assistants that answer questions about your product data.
Help analysts search and synthesize information across large document collections.
Every plan starts with a 7-day free trial. No credit card required.
For side projects and small apps.
For teams shipping to production.
For growing products at scale.
For high-volume production workloads.
Stop building retrieval infrastructure. Start shipping AI features.