Engine
Search API and library — BM25, vectors, hybrid, and geo with pluggable snapshot storage.
Anvesh
Search you can own — crawl, index, and query.
Lightweight Node.js search stack — crawl, index, and search with BM25, vectors, hybrid, and geo — plus a Hub control plane with RBAC.
Product
Real search without the heavy cluster
Anvesh is search you can own: engine, spider, indexer, and Hub in one open-source monorepo. Run keyword, semantic, hybrid, and geo search without standing up an Elasticsearch cluster or paying a hosted search SaaS per query. Adapters speak Elasticsearch, OpenSearch, and Solr when you outgrow native storage — and Vaakly polishes API summary messages by default.
- Full-text (BM25), vectors, hybrid, and geo in one engine
- Crawl → index → search in a single Node.js monorepo
- Hub UI with RBAC for operators across engine, spider, and indexer
- Adapters to Elasticsearch / OpenSearch / Solr with one query shape
- Vaakly plugin for corrected API summary messages
- Local-first — npm start; containers optional for ECS/EKS
Stack
Engine, spider, indexer, Hub
Most products are only an engine or only a crawler. Anvesh ships the pipeline as one Node.js monorepo.
Spider
Site crawler with role-aware discovery (guest / logged-in) for public and member content.
Indexer
Bulk load JSON/JSONL into indexes — the path from crawl or export to searchable docs.
Hub
Optional control plane — instances, indexes, jobs, and admin / operator / viewer RBAC.
Flow
From crawl to query
Spider discovers pages. Indexer bulk-writes documents. Engine serves keyword, semantic, hybrid, and geo search. Hub is optional for search — and the control plane when configured.
01
Crawl
Spider discovers public and role-gated pages.
02
Index
Indexer loads JSON/JSONL into engine indexes.
03
Search
Engine answers BM25, vector, hybrid, and geo queries.
04
Operate
Hub manages instances, jobs, and RBAC when you need it.
Packages
What ships in the monorepo
anvesh-engine
Search API (BM25, vectors, geo)
anvesh-hub
Control plane + modern UI
anvesh-indexer
Bulk indexer worker
anvesh-spider
Site crawler worker
anvesh-setup
Easy local installer
anvesh-plugins
LLM-tool-style plugin host
Fit
When we recommend Anvesh
Excellent for product and site search you can run yourself. Not a drop-in for planet-scale Elastic clusters.
- Product, docs, and help-center search you can run yourself
- Internal knowledge bases and RAG corpora under low millions of docs
- Crawl-backed site search with role-based content
- Teams on Node.js that want ownership without a second ops team
Next steps
Continue exploring
Open source
Registry, docs, and source
Install from the registry, read the docs, or browse the source.
Also from VaagaTech
Related products
Developer tools
Snapline
Declarative snapshot and reconciliation testing for APIs, databases, and async workflows — with an optional Hub for shared test history.
E-commerce
Vastra Platform
Create and host online stores with a lean, low-cost footprint — built for merchants and growing brands.
Language polish
Vaakly
Correct grammar and spelling in English you already have — offline, lightweight, no invented content. Built for API messages, logs, notes, and search snippets.
Need search without the heavy ops?
Tell us about your corpus, crawl needs, and whether you want Hub as the control plane.