New term mining

Every six hours we sweep five free sources — GitHub, Hacker News, Hugging Face, npm and Google News — extract candidate names, deduplicate them across runs, and score how new, corroborated and searchable each one is.

tracked candidates

observations stored

new last run

Ready to build

Discovery pipeline

Candidates flow through stage gates. Only the top shortlist pays for deep SEO and domain checks — the expensive step stays cheap by design.

0Observedseen at least once

0Corroborated2+ independent sources

0SEO validateddemand evidence found

0Ready to buildvalidated, score 0.70+

Source status — last run

Updated Run

How scoring works

Sub-scores are normalized 0–1 per meaning; raw cross-platform metrics (stars vs likes vs mentions) are never compared directly. Established terms with a dedicated Wikipedia entry are archived, not promoted.

30%

Corroboration

25%

Growth

20%

Freshness

15%

Demand

10%

Competition

Top opportunities

Validated candidates ranked by total score, with the evidence trail behind each one.

VideoDub AI

productFirst seen Oct 13 2026github, gnews, npm

validated0.78

GitHub repo streamtools/video-dub-ai (268 stars); 4 news mention(s); npm package video-dub-ai

Score breakdown

Corroboration 30%

0.300

Growth 25%

0.188

Freshness 20%

0.196

Demand 15%

0.075

Competition 10%

0.015

Why it scores this way

sources:github+gnews+npm
sources 2->3
first seen 20h ago
autocomplete:4

videodubai.com — availablevideodubai.io — taken

PromptForge Studio

repoFirst seen Oct 12 2026github, hackernews

validated0.76

GitHub repo acme-labs/promptforge-studio (412 stars); HN discussion: Show HN: PromptForge Studio — version control for LLM prompts

Score breakdown

Corroboration 30%

0.201

Growth 25%

0.188

Freshness 20%

0.186

Demand 15%

0.094

Competition 10%

0.010

Why it scores this way

sources:github+hackernews
sources 1->2
first seen 45h ago
autocomplete:5

promptforgestudio.com — availablepromptforgestudio.io — taken

AgentMemoryKit

repoFirst seen Oct 13 2026github, hackernews

corroborated0.62

GitHub repo forgetful-ai/agent-memory-kit (96 stars); HN discussion: AgentMemoryKit — long-term memory for AI agents

Score breakdown

Corroboration 30%

0.201

Growth 25%

0.188

Freshness 20%

0.198

Demand 15%

0.000

Competition 10%

0.000

Why it scores this way

sources:github+hackernews
sources 1->2
first seen 6h ago

RagEval Bench

modelFirst seen Oct 13 2026huggingface, gnews

corroborated0.55

Hugging Face model org-rag/rag-eval-bench; 2 news mention(s)

Score breakdown

Corroboration 30%

0.201

Growth 25%

0.125

Freshness 20%

0.196

Demand 15%

0.000

Competition 10%

0.000

Why it scores this way

sources:gnews+huggingface
new this run
first seen 10h ago

All tracked candidates

Candidate	Type	First seen	Sources	Stage	Score
VideoDub AI	product	Oct 13 2026	github, gnews, npm	validated	0.78
PromptForge Studio	repo	Oct 12 2026	github, hackernews	validated	0.76
AgentMemoryKit	repo	Oct 13 2026	github, hackernews	corroborated	0.62
RagEval Bench	model	Oct 13 2026	huggingface, gnews	corroborated	0.55
SitemapWatcher Pro	package	Oct 14 2026	npm	observed	0.43
VoiceAgentRouter	repo	Oct 14 2026	github	observed	0.43

Run

How scoring works