> For the complete documentation index, see [llms.txt](https://ai-os-and-trend-finder.gitbook.io/ai-os-and-trend-finder-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ai-os-and-trend-finder.gitbook.io/ai-os-and-trend-finder-docs/docs/research/parallel-projects.md). # Parallel Open-Source Projects For Trend Finder Last reviewed: 2026-05-24 This is a point-in-time external research snapshot for the Trend Finder extension. It was not externally refreshed during the 2026-05-26 documentation audit. It is not an implementation inventory for AI OS. For implemented behavior, use the root [README](/ai-os-and-trend-finder-docs/readme.md), [Apify setup](/ai-os-and-trend-finder-docs/docs/apify.md), and [hackathon submission guide](/ai-os-and-trend-finder-docs/docs/hackathon/hackathon-submission.md). **None of these sources were used in the end, but are left in the project for future consideration.** Trend Finder today is a local AI OS extension that ranks AI trend topics with source evidence, score breakdowns, creator angles, hidden gems, generated watchlist rows, source health, editable Creator Lens run control, a Brief view, and Engine Replay Replay/Reference modes. It has a direct Hacker News adapter plus reviewed Apify source declarations for GitHub public repository metadata, Reddit metadata, arXiv metadata, Product Hunt launches, YouTube video metadata, and RSS/news feeds when configured by the operator. ## Ranking Criteria Projects were ranked on 2026-05-24 for relevance to the current Trend Finder extension at that time, not generic popularity. The strongest matches: * collect from sources that overlap Trend Finder's roles: developer, discussion, research, launch, creator, and news; * turn noisy feeds into ranked topics, reports, or briefings; * preserve source links, provenance, scoring, or explainability; * support local/self-hosted operation or transparent automation; * have an explicit open-source license suitable for code reuse review. Public but unlicensed or non-commercially licensed repositories are listed as research references later in this document, not as safe code-reuse sources. ## Top 5 | Rank | Project | License | Why It Matters For Trend Finder | | ---- | ----------------------------------------------------------------------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | 1 | [Thysrael/Horizon](https://github.com/Thysrael/Horizon) | MIT | Closest architecture match for a self-hosted radar: configurable HN, RSS, Reddit, Telegram, X/Twitter, GitHub, and OpenBB sources; fetch, deduplicate, score, enrich, summarize, then publish daily briefings. | | 2 | [sansan0/TrendRadar](https://github.com/sansan0/TrendRadar) | GPL-3.0 | Mature broad trend monitor with multi-platform hot lists, RSS, AI filtering, AI analysis, schedules, local/cloud self-hosting, and push channels. Strong reference for source health, filtering, and alert ergonomics. | | 3 | [liyedanpdx/reddit-ai-trends](https://github.com/liyedanpdx/reddit-ai-trends) | MIT | Direct AI-trend analogue for one source family: scans AI Reddit communities, analyzes posts, tracks daily rankings, and publishes English/Chinese reports. Useful for discussion-source normalization and longitudinal community trend language. | | 4 | [hoodini/yuv-ai-trends](https://github.com/hoodini/yuv-ai-trends) | MIT | AI/ML trend dashboard focused on GitHub and Hugging Face, with smart ranking, summaries, RSS/JSON feeds, and local API-key storage. Good product reference for developer-source trend cards and privacy framing. | | 5 | [liyown/ai-trend-publish](https://github.com/liyown/ai-trend-publish) | MIT | TypeScript/Deno content workflow that crawls web/RSS/Twitter sources, uses AI for ranking and summaries, renders preview HTML, and exposes a dashboard. Relevant to Trend Finder's creator handoff and future report/export work, even though it is publishing-oriented. | ## Project Notes ### 1. Thysrael/Horizon Horizon is the best parallel for Trend Finder's source-to-briefing pipeline. It models the same core loop Trend Finder needs to keep tightening: * configured source set; * concurrent fetch; * deduplication across platforms; * AI scoring and thresholding; * enrichment with background context and community discussion; * daily Markdown/GitHub Pages, email, webhook, and MCP outputs. Trend Finder should not copy Horizon's broader personal-news posture. The useful reference is the pipeline shape: source adapters feed normalized items, repeated stories are merged, scoring stays configurable, and the final briefing remains tied to evidence. ### 2. sansan0/TrendRadar TrendRadar is less AI-specific than Trend Finder, but it is a strong operational benchmark. It already handles ongoing monitoring, schedule windows, keyword and AI-interest filtering, fallback behavior, display regions, and many push channels. For Trend Finder, the most relevant ideas are: * make source and filter state visible instead of hiding failures; * support operator-owned schedules before adding any hosted service; * separate keyword filtering from AI-assisted relevance filtering; * treat notifications as downstream outputs, not the primary intelligence model. Because it is GPL-3.0, any code reuse would require careful license review. ### 3. liyedanpdx/reddit-ai-trends reddit-ai-trends is narrow but very close to Trend Finder's discussion-source role. It focuses on AI-related Reddit communities, daily hot-topic ranking, trend reports, community discussion summaries, multilingual output, and cached enrichment. The strongest takeaways for Trend Finder are: * a single high-quality community source can produce useful trend language; * daily reports need historical context, not just today's top posts; * discussion sources benefit from bot/spam handling, comment summarization, and repeat-topic tracking; * multilingual reporting is useful but should be downstream of the evidence model, not mixed into normalization. ### 4. hoodini/yuv-ai-trends yuv-ai-trends overlaps Trend Finder's developer and research roles. It is aimed at AI developers and pulls GitHub and Hugging Face signals into a dashboard with ranking, summaries, trend indicators, RSS/JSON feeds, and a local-first API-key story. The useful reference is its narrow product framing: "AI/ML trends for developers" is easier to trust than a generic news feed. Trend Finder's own equivalent should remain "AI topics for creators and automators", with score factors and creator angles visible beside the evidence. ### 5. liyown/ai-trend-publish TrendPublish is not a competing trend dashboard; it is a content-production system. It matters because Trend Finder's Brief view is also a creator handoff, and the deferred generated HTML report will need dry-run, preview, template, source, and secret-boundary discipline. Useful patterns: * keep provider credentials separate from feature configuration; * support dry-run preview before publishing or exporting; * make data-source groups explicit; * use AI for sorting, summary, title, and rendering only after source collection is clear; * keep generated artifacts out of source control unless they are safe examples. ## Strong References With License Caveats These projects are highly relevant as product or source-model references, but they were not counted in the top five because the visible repository metadata does not provide a straightforward open-source code-reuse license. | Project | Relevance | Caveat | | ------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------- | | [duanyytop/agents-radar](https://github.com/duanyytop/agents-radar) | Probably the closest source-model benchmark: GitHub repos, Claude Code Skills, GitHub Trending, Hacker News, Product Hunt, arXiv, Hugging Face, Dev.to, Lobste.rs, Anthropic, and OpenAI; daily, weekly, monthly reports; GitHub Pages, RSS, and MCP access. | No explicit license detected on 2026-05-24. Use for product comparison and source taxonomy, not code reuse unless permission or a license appears. | | [BuilderPulse/BuilderPulse](https://github.com/BuilderPulse/BuilderPulse) | Strong creator-angle benchmark: cross-validates public signals from Hacker News, GitHub Trending, Product Hunt, Hugging Face, Google Trends, Reddit, Indie Hackers, Lobsters, and DEV Community into concrete builder ideas. | Repository content is marked CC BY-NC 4.0 / non-commercial. Treat as positioning inspiration, not reusable product code or commercial data. | | [LearnPrompt/ai-news-radar](https://github.com/LearnPrompt/ai-news-radar) | Relevant lightweight AI/news radar pattern with GitHub Actions, RSS-style source automation, and web UI. | No explicit license detected on 2026-05-24. | ## Implications For Trend Finder The external landscape is crowded around feeds and digests. Trend Finder's defensible difference should stay focused: * topic-level AI trend evidence, not a general news inbox; * transparent score factors: momentum, novelty, evidence strength, source diversity, niche fit, and creator potential; * source health and provenance labels for live, fixture/demo, degraded, blocked, fallback, and unknown states; * editable Creator Lens run control that stays local to AI OS; * generated watchlist, movement, prediction, and retro rows rather than manual saved items; * creator angles and Brief handoff grounded in public evidence links; * local-first operation through AI OS, with private runtime data and credentials kept out of git. Future work should borrow concepts, not claims: better deduplication, richer source history, configurable thresholds, source-specific freshness windows, and optional report export are all consistent with Trend Finder's current direction. Hosted operation, unrestricted source collection, and automated publishing remain outside the implemented MVP unless a later spec explicitly adds them. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://ai-os-and-trend-finder.gitbook.io/ai-os-and-trend-finder-docs/docs/research/parallel-projects.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.