Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Personalized paper recommendation for OpenClaw / Feishu, powered by AMiner + arXiv + LLMs.
| Date | Stars |
|---|---|
| 2026-07-31 | 475 |
| 2026-08-04 | 475 |
| 2026-08-06 | 475 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<div align="center">
# 📚 aminer-rec
**Stop drowning in arXiv. Start reading what actually matters to *you*.**
A personalized paper-recommendation engine that turns *one sentence* about your research into a ranked, summarized reading list — built on AMiner + arXiv + LLMs.
</div>
---
## ✨ Why you'll like it
Picture this: you walk in Monday morning, type
> *"I work on multimodal agents and tool use"*
…and a minute later you get a clean **ranked shortlist** of recent papers, each with a **plain-language summary** and a **one-line reason** for why it landed in your feed. No more 200-tab arXiv sessions. No more "did I miss the important one?".
`aminer-rec` does the whole loop for you:
| You give it | It gives you back |
|---|---|
| a sentence, a topic list, **or** a scholar name | a focused list of recent, relevant papers |
| your AMiner scholar id / seed papers | a *profile-aware* ranking tuned to *your* taste |
| `--language-sort en` / `--start-year 2024` | filtering by language and year |
## 🎯 Two ways to start, one unified pipeline
Pick whichever feels lazier:
- 🧠 **Topic bootstrap** — just describe what you do in plain language.
> `--free-text "I work on multimodal agents and tool use"`
- 🎓 **Scholar bootstrap** — start from an `aminer_user_id`, a name + org, or a few seed paper titles. The pipeline builds a `ResearchProfile` from your real publication history and uses it to rank.
If you are not sure which input to use, pick from this table:
| Your situation | Recommended input | Example |
|---|---|---|
| You only have a rough direction | `--free-text` | `--free-text "I work on multimodal agents and tool use"` |
| You already know the keywords | `--topics` | `--topics "LLM reasoning, tool use"` |
| You want a scholar-shaped profile | `--scholar-name` + `--scholar-org` | `--scholar-name "Jie Tang" --scholar-org "Tsinghua University"` |
| You know the AMiner scholar id | `--aminer-user-id` | `--aminer-user-id 696259801cb939bc391d3a37` |
| You want to start from seed papers | `--paper-title` | `--paper-title "OAG-Bench" --paper-title "RPC-Bench"` |
Both paths collapse into a single `ResearchProfile`, then flow through the same pipeline:
```mermaid
flowchart LR
A1["topics / free text"] --> B["build ResearchProfile"]
A2["scholar / id / seed papers"] --> B
B --> C["arXiv retrieval"]
C --> D["AMiner enrichment"]
D --> E["profile-aware ranking"]
E --> F["LLM summary + recommendation reason"]
F --> G["Markdown · JSON"]
F --> H["Feishu cards\n(optional)"]
```
## 👀 What the output looks like
Run it once, and `outputs_cli/recommendation.md` is a clean reading list — every paper carries a summary and a one-line reason it was picked for you. Here's the shape of a single entry (the renderer emits Chinese labels by default):
```markdown
为你推荐 5 篇相关论文（研究方向：multimodal agents / tool use）
---
**1. [ToolGen: Unified Tool Retrieval and Generation...](https://www.aminer.cn/pub/5f...)**
年份：2024 | 关键词：LLM · tool use · agents
作者：X. Y. Zhang、A. B. Li、C. Wang et al.
推荐理由：与你的 "tool use" 方向高度吻合，提出统一工具检索与生成的框架，正好补你最近关注的能力缺口。
本文提出 ToolGen，将工具的检索与调用统一进一个生成式框架……（摘要正文）
```
The companion `recommendation_result.json` carries the full structured payload — `title`, `keywords`, `authors`, `famous_authors`, `summary`, `recommendation_reason`, `aminer_paper_url`, and more — so you can feed it into anything downstream.
To preview the output shape before configuring tokens, open [examples/topic_bootstrap.recommendation.md](examples/topic_bootstrap.recommendation.md) and [examples/topic_bootstrap.recommendation_result.json](examples/topic_bootstrap.recommendation_result.json). These are static examples, not live retrieval results.
## 🚀 Excerpt of 9,662 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:d8d08fefd1128ec8, llm:description: 'Personalized paper recommendation for OpenClaw / Feishu, powered by AMiner + arXiv + LLMs.'
matched fp:d8d08fefd1128ec8, llm:description: 'Personalized paper recommendation for OpenClaw / Feishu, powered by AMiner + arXiv + LLMs.'
matched fp:d8d08fefd1128ec8, llm:description: 'Personalized paper recommendation for OpenClaw / Feishu, powered by AMiner + arXiv + LLMs.'