Cloud-Native Search & Vectors
for the Modern Stack

Sub-millisecond hybrid BM25 + vector search, non-AI visual/OCR feature extraction, GitOps declarative schema sync, and 3-tier concurrency throttling — all in an ultra-lightweight Node.js engine.

🚀 Get Started Try Live Demo

<1ms

Query Latency (p99)

85%

Lower TCO

~70MB

Memory Footprint

Documents Lost (DLQ)

Capabilities

Everything you need for enterprise search

Full-text BM25, dense vector retrieval, non-AI OCR & visual extraction, GitOps declarative schemas, and multi-layered concurrency protections.

⚡ Hybrid Search

BM25 full-text + dense vector semantic search with reciprocal rank fusion (RRF) and semantic synonym highlights.

🎯 Field Projections

Granular field selection, inclusion/exclusion syntax ({ title: 1, body: 0 }), dot-notation nested paths, and document ID control.

⏰ Spider Scheduler

Automated 5-part cron and interval website crawlers with on-demand "Run Now" triggers and run history tracking.

🔄 Multi-Source Scanner (CDC)

Incremental change data capture for local/network folders (SHA-256 fingerprinting), streaming files, and SQL databases.

🎨 Non-AI Visual & OCR

Local pure-CPU OCR text recognition, textile color palette matching, and motif/pattern edge descriptors without external AI models.

⚙️ Config-as-Code & GitOps

Declarative schema evolution and drift detection via anvesh plan, anvesh apply, or Terraform without restarting pods.

🛡️ 3-Tier Protective Throttling

Rate limiting (RPM), in-flight concurrent query slot caps, and proactive 85% heap backpressure guards against runaway loads.

💾 Distributed State & Storage

Zero pod-local SQLite; state shared across nodes via internal system indexes and tiered object storage (S3/OCI).

🔒 Zero-Drop DLQ

Zero-document-loss guarantee. Every failed write is captured in an isolated DLQ with automated retries and one-click replay.

Developer First

Programmatic SDK & CLI Tooling

📦 TypeScript SDK

import { AnveshClient } from "@vaagatech/anvesh-sdk";

const client = new AnveshClient({
  baseUrl: "https://api.search.example.com",
  apiKey: process.env.ANVESH_API_KEY,
});

const results = await client.search("products", {
  q: "festive silk saree with elephant",
  mode: "hybrid",
});

💻 Anvesh CLI & GitOps

# Initialize declarative config
anvesh init

# Detect drift vs live cluster
anvesh plan -f anvesh.config.json

# Apply schema without restarts
anvesh apply -f anvesh.config.json

# Run instant hybrid search
anvesh search products -q "saree"

Cloud-Native Search & Vectors for the Modern Stack