overview glossary

Glossary

Term	Definition
Adapter	A Python class that wraps a single search engine's API or scrape interface. Each adapter lives in one file in engines/ and is registered via @register_engine.
AdapterResponse	The canonical return type for every adapter's search() method. Contains a list of SearchResult, a status enum, optional error message, and measured latency.
Backpressure	A feedback mechanism where the rate limiter signals upstream clients (or the orchestrator) to slow down when the system approaches capacity limits. Implemented via Valkey sliding window counters.
Category routing	Filtering of engines by SearXNG-compatible category tags. Clients can pass ?categories=science,news to restrict results to engines declaring those categories.
EngineAdapter	Abstract base class for API-based engines (Brave, Wikipedia, etc.)
EngineStatsTracker	Subsystem in slopsearx/stats.py that records per-engine quality telemetry (success/failure counts, latency histograms) in Valkey after each dispatch. Enables adaptive engine selection and observability.
PresenceRanker	V1 ranking strategy that boosts results appearing in multiple engines. Provides breadth at the cost of precision.
ProxyPool	Subsystem in slopsearx/proxypool.py that manages a configurable list of proxy URLs for scrape-based adapters. Provides round-robin proxy rotation to evade IP bans and CAPTCHA walls.
QueryRouter	Subsystem in slopsearx/router.py that analyzes incoming queries and selects only relevant engines based on topic matching, category filters, and explicit engine selections. Reduces unnecessary dispatch to irrelevant engines.
Ranker	Pluggable interface for ranking strategies. V1 ships with PresenceRanker; V2 plans include WeightedFusionRanker with per-engine trust scores.
ScrapeAdapter	Abstract base class for HTML-scrape engines (DuckDuckGo, Google). Sends HTTP requests with stealth headers and parses HTML with lxml.
Scrape engine	An engine that sends HTTP GET/POST requests to a search engine's HTML page and parses the response. No headless browser required.
SearchResult	Internal normalized result dataclass decoupled from any output format. Contains url, title, content, engine, score, position, and metadata.
SearXNG	The open-source meta search engine that SlopSearX is designed to replace. SlopSearX preserves the SearXNG JSON response contract for backward compatibility.
Security engines	Engines focused on vulnerability research, threat intelligence, and security scanning. Examples: Shodan, Censys, VirusTotal, NVD, Exploit-DB.
SuggestionService	Subsystem in slopsearx/suggest.py that runs as a background task during search queries, fetching suggestions from engine suggest APIs to populate the suggestions response field.
Threat-intel engines	A subset of security engines specializing in threat intelligence data: IP reputation (AbuseIPDB, GreyNoise), credential leaks (HIBP, DeHashed), dark web intelligence (IntelX), and exploit data (Exploit-DB, VulnCheck).
Tier 1 engines	Reliable API-based engines (Brave, Wikipedia) that form the reliability backbone
Tier 2 engines	Best-effort scrape-based engines (DuckDuckGo, Google) with no SLA
Topic-based routing	The QueryRouter's strategy for matching query content to engine categories using keyword heuristics, reducing the set of dispatched engines to those likely to return relevant results.
Valkey	Redis-compatible in-memory data store used for caching, distributed rate limiting, and engine quality statistics.
Valkey sliding window	A distributed rate limiting algorithm that uses Valkey sorted sets to track request timestamps within a moving time window. Correct across 50+ replicas.
GroktoCrawl	The larger AI agent stack that SlopSearX integrates into, replacing its SearXNG component
OpenMetrics	The standard format for exposing metrics (Prometheus text format). SlopSearX's /metrics endpoint emits OpenMetrics without the prometheus-client library.
ConfigMap	Kubernetes resource for mounting configuration files. The optional config.yaml is typically provided via ConfigMap.
HPA	Horizontal Pod Autoscaler — Kubernetes resource that automatically scales replicas based on CPU utilization.