Objective package recommendations, scored from real signals.
lurq is a continuously-updated, evidence-scored index of JS/TS frameworks and libraries - so your coding agent recommends dependencies that are real, healthy, and current, not frozen in training data.
Works inside your coding agent
Claude Code
Cursor
Windsurf
GitHub Copilot
VS Code
CodexClaude CodeCursorWindsurfGitHub CopilotVS CodeCodexClaude CodeCursorWindsurfGitHub CopilotVS CodeCodexClaude CodeCursorWindsurfGitHub CopilotVS CodeCodex$ lurq recommend "a form library for react"
dataAsOf 2026-06-22 · scored from npm · github · deps.dev
Recommendations your agent can actually trust.
lurq ingests public signals from npm, GitHub, and deps.dev, scores every package for health and confidence, and exposes the result to your coding agent over MCP and the CLI.
Ask in plain language; get back a short, ranked, evidence-backed list - each entry carrying a confidence label and a freshness timestamp.
Built to outperform the alternatives.
Most tools hand your agent stale or popularity-ranked guesses. lurq is fresh, objective, and verifiable.
Fresh, not frozen in training data
Scores are recomputed from live npm, GitHub, and deps.dev signals – so your agent sees packages and versions released long after the model's cutoff.
Objective scoring from real signals
Health and confidence are derived from downloads, release cadence, maintenance, and security data - never hand-written opinions or popularity alone.
An anti-hallucination guard
verify catches abandoned dependencies before your agent ever installs a package
Agent-native by design
MCP server, CLI, and installable skill with compact, token-aware responses. Works inside the tools your agent already runs in.
Stack-aware recommendations
compare ranks alternatives head-to-head and diagram sketches a reference architecture for the stack you've chosen.
Loved by builders and their agents.
“My agent stopped suggesting packages that were deprecated a year ago. That alone paid for itself.”
“The verify tool caught a typosquatted dependency before it ever hit our lockfile.”
“Finally, recommendations backed by signals instead of vibes. The confidence labels are clutch.”
“compare turned a two-hour evaluation into a thirty-second answer I could trust.”
“My agent stopped suggesting packages that were deprecated a year ago. That alone paid for itself.”
“The verify tool caught a typosquatted dependency before it ever hit our lockfile.”
“Finally, recommendations backed by signals instead of vibes. The confidence labels are clutch.”
“compare turned a two-hour evaluation into a thirty-second answer I could trust.”
“My agent stopped suggesting packages that were deprecated a year ago. That alone paid for itself.”
“The verify tool caught a typosquatted dependency before it ever hit our lockfile.”
“Finally, recommendations backed by signals instead of vibes. The confidence labels are clutch.”
“compare turned a two-hour evaluation into a thirty-second answer I could trust.”
“My agent stopped suggesting packages that were deprecated a year ago. That alone paid for itself.”
“The verify tool caught a typosquatted dependency before it ever hit our lockfile.”
“Finally, recommendations backed by signals instead of vibes. The confidence labels are clutch.”
“compare turned a two-hour evaluation into a thirty-second answer I could trust.”
“Dropping it into Cursor took one command. It just shows up where I already work.”
“The freshness is the whole point. New libraries my model had never heard of, scored and ready.”
“Token-aware responses mean I'm not burning context on giant package dumps. Clean.”
“The diagram tool sketched a sane reference stack I could hand straight to my team.”
“Dropping it into Cursor took one command. It just shows up where I already work.”
“The freshness is the whole point. New libraries my model had never heard of, scored and ready.”
“Token-aware responses mean I'm not burning context on giant package dumps. Clean.”
“The diagram tool sketched a sane reference stack I could hand straight to my team.”
“Dropping it into Cursor took one command. It just shows up where I already work.”
“The freshness is the whole point. New libraries my model had never heard of, scored and ready.”
“Token-aware responses mean I'm not burning context on giant package dumps. Clean.”
“The diagram tool sketched a sane reference stack I could hand straight to my team.”
“Dropping it into Cursor took one command. It just shows up where I already work.”
“The freshness is the whole point. New libraries my model had never heard of, scored and ready.”
“Token-aware responses mean I'm not burning context on giant package dumps. Clean.”
“The diagram tool sketched a sane reference stack I could hand straight to my team.”
Frequently asked questions
Still have questions?
Reach out about the project, partnerships, or anything else – we'll get back to you.