CS Research Intelligence Platform — 100,000+ papers, scored, ranked, and searchable.
Stop skimming paper lists. ResearchScope scores papers by impact, surfaces research gaps, recommends venues, and tracks who's driving the frontier — updated daily.
ResearchScope is an open research intelligence platform for computer science and AI. A daily GitHub Actions pipeline fetches papers from 6 data sources (arXiv, OpenAlex, ACL Anthology, OpenReview, PMLR, CVF, Semantic Scholar), enriches them with multi-signal scores, detects research gaps, and syncs to two backends:
- Railway PostgreSQL — full dataset, powers the REST API with full-text search and live browser queries
- Hugging Face Hub — public JSONL dataset for LLM training
The frontend is a static site on GitHub Pages backed by a FastAPI REST API on Railway.
👉 Open ResearchScope · 📖 API Docs
The researchscope package ships a lightweight command-line tool for searching, saving, and analysing papers locally. It requires Python 3.10+.
pip install researchscopeThis installs the researchscope command. Run it with no arguments to see all commands and options:
researchscope --helpSearch arXiv (or Semantic Scholar) and print ranked results:
# Top 10 arXiv results for a query
researchscope search "graph neural networks"
# Limit results and save them to a local store for later analysis
researchscope search "diffusion models" --limit 25 --save
# Use Semantic Scholar as the source
researchscope search "retrieval augmented generation" --source semantic_scholarList the papers you've saved locally:
researchscope list-papers # all saved papers, ranked
researchscope list-papers --limit 20Gaps — surface the least-covered topics across your saved papers:
researchscope gaps # top 10 gap topics
researchscope gaps --top-n 25Note:
search --savepersists results to a local store first;list-papersandgapsoperate on those saved papers. Run a save before using them.
The CLI is a self-contained prototype. The full ResearchScope platform — 100K+ papers, venue recommenders, and the web dashboard — lives at the links above and in
src/.
| Date | Highlight |
|---|---|
| Jul 2026 | Three-theme interface — pick Atelier Zero (warm editorial), Industrial Brutalist (high-contrast), or Field Notes (tactile) from the nav; selection persists across pages and reloads with no flash. Rebuilt responsive navigation + modular CSS architecture |
| Jun 2026 | OpenReview Acceptance Tiers — oral/spotlight/poster signals captured for ICLR, NeurIPS, ICML & COLM; oral/spotlight boost paper scores and show as badges. Coverage extended through ICLR 2026, NeurIPS 2025, ICML 2025 |
| Jun 2026 | Journal Recommender — paste title + abstract to match against 20 Q1 journals (JMLR, TPAMI, Nature MI, CSUR…) with impact factor, review timeline, and open access info |
| Jun 2026 | FastAPI Backend on Railway — full REST API with JWT auth, favourites, PostgreSQL full-text search (100K+ papers). User accounts synced across devices |
| Jun 2026 | OpenAlex Integration — 250M+ work catalogue added as a data source, covering ML/NLP/CV/IR concept groups |
| Jun 2026 | HuggingFace Training Dataset — kishormorol/researchscope-papers auto-pushed after every pipeline run: raw metadata JSONL + instruction-tuning pairs |
| Jun 2026 | 20 Q1 Journals — JMLR, TMLR, TACL, TPAMI, IJCV, AIJ, TNNLS, Nature MI, CSUR, TIP, MLJ, TKDE, DAMI, NN, PR, CL, IPM, JACM, NatComms, TOIS |
| May 2026 | Complete A Coverage* — AAAI, IJCAI, CHI, SIGIR, WWW, KDD, WSDM, SIGMOD, ICSE bulk-fetched via Semantic Scholar |
| Apr 2026 | Conference Recommender — TF-IDF venue matching with deadline info and reviewer expectations |
| Apr 2026 | CiteLens Integration — one-click citation analysis for any arXiv paper |
| Feature | Description |
|---|---|
| 📄 100K+ papers | Scored by recency, venue rank, acceptance tier (oral/spotlight), novelty, author prestige, and citation quality |
| 🎓 A Conference coverage* | NeurIPS, ICML, ICLR, CVPR, ACL, EMNLP, AAAI, IJCAI, CHI, SIGIR, WWW, KDD and more |
| 📖 20 Q1 Journals | JMLR, TMLR, TACL, TPAMI, Nature MI, and 15 more — with IF, review time, OA status |
| 🎯 Venue Recommenders | Conference + Journal recommenders: paste abstract → ranked matches with expectations |
| ⏰ Deadline tracker | Live countdowns to A*/A conference deadlines across 10 CS areas — abstract, paper, and notification dates |
| 🔍 Full-text search | PostgreSQL tsvector search across 100K+ papers via Railway API |
| 👤 User accounts | JWT auth, favourites synced across devices via Railway backend |
| 💬 Chat with Paper | Authenticated split-screen PDF chat with page citations and cross-device history (provider-configured) |
| 🕳 Research gaps | 3-layer extraction: explicit, pattern-detected, and starter ideas |
| 👩🔬 Author intelligence | 5,000+ researchers ranked by momentum score |
| 🤗 LLM training data | papers.jsonl + instruct.jsonl on HuggingFace Hub |
| 🔗 CiteLens | One-click handoff to citation analysis for arXiv papers |
| 🎨 Three themes | Atelier Zero, Industrial Brutalist, and Field Notes — switch from the nav, persists with no flash |
Live countdowns to top CS conference deadlines — filter by area, track abstract vs. paper deadlines, and review past cycles.
┌──────────────────────────────────────────────────────────────────────┐
│ GitHub Actions │
│ daily (weekdays) + monthly (conference sync) │
│ │
│ src/pipeline.py — 11 stages │
│ ├── connectors/ arXiv · OpenAlex · ACL · OpenReview · PMLR │
│ │ CVF · Semantic Scholar │
│ ├── dedup/ Jaccard title-similarity dedup │
│ ├── tagging/ 80+ topic tags + paper_type │
│ ├── difficulty/ L1–L4 reading level │
│ ├── scoring/ 4 scores + author momentum │
│ ├── content/ summaries, key contributions, why-it-matters │
│ ├── clustering/ topic grouping │
│ ├── gaps/ 3-layer research gap extraction │
│ ├── aggregation/ author, lab, university profiles │
│ └── sitegen/ → site/data/*.json │
│ → Railway PostgreSQL upsert (API backend) │
│ → HuggingFace Hub push (JSONL dataset) │
└────────┬─────────────────────────┬──────────────────────────────────┘
│ commits + deploys │ syncs │ pushes
▼ ▼ ▼
GitHub Pages Railway PostgreSQL HuggingFace Hub
(static site) FastAPI REST API researchscope-papers
/papers /search papers.jsonl
/auth /favourites instruct.jsonl
/docs (Swagger UI)
▲ ▲
│ reads static JSON │ Railway API (→ static JSON fallback)
└─────────────────────────┘
Frontend (browser)
The REST API is live at https://researchscope-production.up.railway.app.
| Method | Endpoint | Description |
|---|---|---|
GET |
/papers |
Paginated papers — filter by venue, year, source_type, tag |
GET |
/papers/conferences |
Conference papers only |
GET |
/papers/journals |
Journal papers only |
GET |
/papers/{id} |
Single paper |
GET |
/search?q=... |
PostgreSQL full-text search |
POST |
/auth/register |
Create account → JWT |
POST |
/auth/login |
Login → JWT |
GET |
/auth/me |
Current user |
GET |
/favourites |
Saved papers (auth required) |
POST |
/favourites/{id} |
Save paper |
DELETE |
/favourites/{id} |
Remove saved paper |
GET/POST |
/papers/{id}/document-status, /prepare |
Inspect or prepare a PDF for grounded chat |
GET/POST |
/chat/sessions |
List or create user-owned paper chats |
POST |
/chat/sessions/{id}/messages |
Stream a cited answer with server-sent events |
POST |
/pipeline/trigger |
Trigger pipeline via GitHub Actions |
Interactive docs: /docs
The paper dataset is published on HuggingFace and auto-updated after every pipeline run.
from datasets import load_dataset
# 100K+ raw paper records (pretraining / RAG)
papers = load_dataset("kishormorol/researchscope-papers",
data_files="data/papers.jsonl", split="train")
# Instruction-tuning pairs (summarize, key contribution, why it matters…)
instruct = load_dataset("kishormorol/researchscope-papers",
data_files="data/instruct.jsonl", split="train")→ huggingface.co/datasets/kishormorol/researchscope-papers
| Source | Content | Frequency |
|---|---|---|
| arXiv (OAI-PMH) | All cs.* preprints — 19 CoRR categories | Daily |
| OpenAlex | 250M+ works — ML/NLP/CV/IR concept groups | Daily |
| ACL Anthology | ACL, EMNLP, NAACL, EACL, COLING, TACL, CL (2020+) | Monthly |
| OpenReview | ICLR (2022–26), NeurIPS (2022–25), ICML (2024–25), COLM (2024–25) — with oral/spotlight/poster acceptance tiers | Monthly |
| PMLR | ICML (2020–25), AISTATS (2021–25), UAI (2021–24) | Monthly |
| CVF | CVPR (2021–25), ICCV (2021+23), ECCV (2020+22+24) | Monthly |
| Semantic Scholar | AAAI, IJCAI, KDD, WWW, SIGIR, WSDM, CHI, SIGMOD, ICSE + journals | Monthly |
.github/workflows/
pipeline.yml # daily arXiv + OpenAlex pipeline
conference-sync.yml # monthly full conference + journal sync
backfill.yml # manual historical backfill
discord-potd.yml # daily Paper of the Day → Discord
backend/ # FastAPI REST API (deployed on Railway)
app/
main.py # FastAPI app with CORS, lifespan
database.py # async SQLAlchemy + asyncpg
models.py # Paper, User, Favourite ORM models
schemas.py # Pydantic v2 schemas
auth.py # JWT + bcrypt
routers/
papers.py # GET /papers /conferences /journals
search.py # GET /search (PostgreSQL full-text)
auth.py # POST /auth/register /login GET /me
favourites.py # GET POST DELETE /favourites
pipeline.py # POST /pipeline/trigger
requirements.txt
railway.toml
src/
pipeline.py # 11-stage orchestrator
connectors/
arxiv_connector.py
openalex_connector.py # NEW — 250M+ OpenAlex works
acl_connector.py
openreview_connector.py
pmlr_connector.py
cvf_connector.py
semantic_scholar_connector.py
storage/
railway_store.py # Railway PostgreSQL upsert
hf_dataset.py # HuggingFace Hub push
sitegen/
generator.py
conference_recommender.py
journal_recommender.py # NEW — 20 Q1 journals
config/
venues.yaml # conferences + journals registry
topics.yaml # topic taxonomy
weights.yaml # tuneable score weights
site/ # static frontend (GitHub Pages)
index.html # homepage
papers.html # paper browser
conferences.html # A* conference papers
journals.html # Q1 journal papers
journal-recommender.html # NEW
conference-recommender.html
topics.html / authors.html / labs.html
gaps.html / digest.html / deadlines.html
search.html / favourites.html / library.html
assets/js/
app.js # shared utilities + dropdown nav
railway-api.js # Railway API data client + auth
tests/ # pytest suite (110+ tests)
# Clone and install
git clone https://github.com/kishormorol/ResearchScope.git
cd ResearchScope
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
# Run the full pipeline locally
python src/pipeline.py
# Conference + journal sync only
python src/pipeline.py --conferences-only
# Run tests
python -m pytest tests/ -v
# Serve the frontend
cd site && python -m http.server 8080
# Serve the frontend with production-style extensionless routes (from the repository root)
cd .. && python scripts/serve_site.py --port 8080
# Run the API locally
cd backend
pip install -r requirements.txt
DATABASE_URL=postgresql://... uvicorn app.main:app --reload
# Alternatively, use DATABASE_URL from the repository-root .env
uvicorn app.main:app --reloadThe local API loads the repository-root .env without overriding variables
already supplied by the shell or deployment environment.
| Variable | Where | Description |
|---|---|---|
RAILWAY_DATABASE_URL |
GH Secret | Public Railway PostgreSQL URL (pipeline sync) |
DATABASE_URL |
Railway Env | Same URL for FastAPI backend |
JWT_SECRET |
Railway Env | Secret key for JWT signing |
HF_TOKEN |
GH Secret | HuggingFace write token for dataset push |
SEMANTIC_SCHOLAR_API_KEY |
GH Secret | Raises S2 rate limit 1→10 req/s |
OPENREVIEW_EMAIL |
GH Secret | For OpenReview authenticated access |
OPENREVIEW_PASSWORD |
GH Secret | For OpenReview authenticated access |
DISCORD_WEBHOOK_URL |
GH Secret | Paper of the Day → Discord |
PIPELINE_SECRET |
Both | Shared secret for /pipeline/trigger endpoint |
GITHUB_TOKEN |
Railway Env | Fine-grained PAT for triggering workflows |
CHAT_ENABLED |
Railway Env | Feature flag; defaults to false |
OPENAI_API_KEY |
Railway Env | Server-side key for Chat with arXiv; never expose it to the browser |
OPENAI_CHAT_MODEL |
Railway Env | Optional model override; defaults to gpt-5.6-terra |
OPENAI_REASONING_EFFORT |
Railway Env | Optional reasoning level; defaults to low for lower cost and latency |
Chat semantic ranking runs inside PostgreSQL. Standard Railway PostgreSQL uses
the compatible REAL[] cosine query; a Railway pgvector database automatically
adds an HNSW index for the default 256-dimension embeddings without rewriting or
deleting existing embedding data.
ResearchScope and CiteLens are companion tools:
ResearchScope ──── "Here's a paper worth reading today"
│
🔍 Analyze citations
▼
CiteLens ────── "Here's who cited it and why it mattered"
│
🔭 Browse topic in ResearchScope
▼
ResearchScope ──── "Discover more papers on this topic"
| Tool | Free | Open source | Daily updates | Gaps | Venue recommender | API | No sign-up |
|---|---|---|---|---|---|---|---|
| ResearchScope | ✅ | ✅ | ✅ | ✅ | ✅ Conference + Journal | ✅ | ✅ |
| Arxiv Sanity | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ |
| Papers With Code | ✅ | ❌ | ✅ | ❌ | ❌ | Partial | ✅ |
| Semantic Scholar | ✅ | ❌ | ✅ | ❌ | ❌ | ✅ | ✅ |
| Elicit | Partial | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| Source | License |
|---|---|
| arXiv | Metadata: CC0 |
| OpenAlex | CC0 |
| ACL Anthology | CC BY 4.0 |
| PMLR | CC BY 4.0 |
| Semantic Scholar | S2 API License |
| OpenReview | Public API |
| CVF | Public access |
ResearchScope stores only bibliographic metadata — no full text or PDFs.
The statement above describes the public dataset. When Chat with Paper is enabled, the authenticated backend may fetch an allowlisted PDF, discard the PDF bytes after processing, and persist page-aware extracted text chunks for grounded answers. User conversations remain private to their account until deleted.
| Contributor | GitHub | Role |
|---|---|---|
| Md Kishor Morol | @kishormorol | Project lead · architecture · pipeline |
| Mohammed Efaz | @WhiteHades | Multi-theme UI · frontend architecture |
| Shadril Hassan | @shadril238 | Topic network graph |
| Saad Chowdhury | @0Sa-ad0 | Contributor |
Want to contribute? See CONTRIBUTING.md.
MIT © 2026 Md Kishor Morol