# metarank/metarank

A low code Machine Learning personalized ranking service for articles, listings, search results, recommendations that boosts user engagement. A friendly Learn-to-Rank engine

Repository: https://github.com/metarank/metarank
Canonical: https://ross.abutalabs.com/products/metarank
Homepage: https://metarank.ai
Language: Scala
License: Apache-2.0
License Family: permissive
Topics: ranking, scala, search, personalization, machine-learning, deep-learning, data-engineering, feature-engineering, feature-extraction, kubernetes, data-science, automl, neural-networks
Last push: 2026-08-26T10:04:03+00:00

## Health v2 (maintenance only)
Score: 88/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 67, longevity 100
- inputs: {"age_days": 2205, "days_push": 7, "days_rel": 8, "gap_med": 427, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2434, forks 108 (observed 2026-08-28T04:06:51.295822+00:00)

## What it is
Metarank is an open-source, low-code Learning-to-Rank service that personalizes search results, listings, and recommendations using real-time user signals like clicks and purchases. It is a stateless, cloud-native service (with state in Redis) that reranks results from existing search or recommendation systems with low latency.

## Use cases
- add learning-to-rank reranking to my existing search engine
- personalize search results based on user clicks and purchases
- build a 'you may also like' recommendation widget for my store
- optimize ranking for maximum CTR
- add real-time session-based personalization to listings
- use LLM bi-encoders and cross-encoders for semantic search reranking
- deploy a scalable low-latency reranking service on Kubernetes

## When to choose
- you already have a search or recommendation engine and want to boost engagement with ML-based reranking
- you need real-time personalization with 10-20ms reranking latency
- you want prebuilt ranking signals (CTR, referer, user-agent, time) without writing custom feature code
- you need a horizontally scalable, cloud-native ranking service

## When to avoid
- you need a full search engine rather than a reranking layer on top of one
- your stack cannot run JVM services or manage a Redis dependency
- you have no user interaction data to train or feed the ranker
- you need a fully managed SaaS rather than self-hosted infrastructure

## Facets
- artifact type: service
- maturity: active
- function: machine-learning, search-engine, rag, llm-inference, api-framework, monitoring
- domain: machine-learning, data-science, e-commerce, web-development
- platform: self-hosted, cloud, jvm
- tags: learning-to-rank, personalization, reranking, scala, feature-engineering, collaborative-filtering, semantic-search, ctr-optimization, search, recommendation, docker, kubernetes

## Member repositories
- metarank/metarank (main) score 88

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:51.295822+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:31:08.900744+00:00, confidence not recorded.
  - readme: https://github.com/metarank/metarank (fetched 2026-08-28T04:06:51.295822+00:00, sha e1872be829f7)
- Data as of 2026-08-30T08:39:29.467469+00:00.