A visual DuckDuckGo query builder built with Streamlit — designed for data professionals. Compose advanced search queries by combining operators, run them across multiple sources, compare results side by side, and export everything to CSV or JSON.
| Feature | Description |
|---|---|
| 🛠️ Visual query builder | Add operator blocks from the sidebar, reorder with ↑ ↓ arrows, edit values |
| 🎯 Query templates | Pre-built stacks for datasets, papers, GitHub repos, API docs, tutorials, and files |
| 💾 Saved queries | Name and save any operator stack, reload it later in one click |
| ⚖️ Side-by-side compare | Build two independent queries and compare results in two columns |
| 🌐 Batch search | Run one query across up to 10 data/research sites simultaneously |
| 📚 Search history | Last 50 searches auto-saved with timestamp and result count; reload any past query |
| ⬇️ Export | Download results as CSV or JSON per query or per batch site |
| 🔎 Domain filter | Filter result cards by domain without re-searching |
| ♻️ Deduplicate | Strip duplicate URLs from result sets |
| 🕐 Date range filter | Restrict results to past day / week / month / year |
| 🌍 Region filter | Target results by country/language |
| Operator | Example | Description |
|---|---|---|
| 🔍 Keyword | python tutorial |
Plain search term |
🌐 site: |
site:github.com |
Restrict to a domain |
💬 "Exact phrase" |
"open source software" |
Words must appear together |
🚫 -Exclude |
-spam |
Remove results with this word |
✅ +Include |
+free |
Force-include a word |
📄 filetype: |
filetype:pdf |
Match a specific file type |
🔤 intitle: |
intitle:login |
Word must be in the page title |
🔗 inurl: |
inurl:admin |
Word must be in the URL |
🔄 related: |
related:nytimes.com |
Find similar websites |
⚡ OR |
cat OR dog |
Match either term |
🌟 * Wildcard |
"how to * python" |
Wildcard inside a quoted phrase |
Templates load a pre-built operator stack. Fill in the keyword block and search immediately.
| Template | Searches on |
|---|---|
| 📊 Find Dataset | Kaggle, Hugging Face, data.gov |
| 📄 Find Research Paper | arXiv, Papers With Code |
| 💻 Find GitHub Repo | GitHub |
| 📁 Find CSV / Excel File | Any site, filetype:csv or xlsx |
| 📖 Find API Docs | Pages with "API" or "documentation" in title |
| 🎓 Find Tutorial | Pages with "tutorial" in title, excluding paid |
| Site | Domain |
|---|---|
| 🤗 Hugging Face | huggingface.co |
| 📊 Kaggle | kaggle.com |
| 📄 Papers With Code | paperswithcode.com |
| 🔬 arXiv | arxiv.org |
| 💻 GitHub | github.com |
| 🏛️ data.gov | data.gov |
| 🌐 UCI ML Repo | archive.ics.uci.edu |
| 🔷 Stack Overflow | stackoverflow.com |
| 📰 Medium | medium.com |
| 📝 Towards Data Science | towardsdatascience.com |
| Label | Code |
|---|---|
| 🌍 Worldwide | wt-wt |
| 🇺🇸 USA | us-en |
| 🇬🇧 UK | uk-en |
| 🇩🇪 Germany | de-de |
| 🇫🇷 France | fr-fr |
| 🇯🇵 Japan | jp-ja |
| 🇧🇷 Brazil | br-pt |
| 🇮🇳 India | in-en |
| 🇦🇺 Australia | au-en |
| 🇨🇦 Canada | ca-en |
- Python 3.10+
- uv (recommended) — or any virtual environment tool
git clone https://github.com/49mg/easy-search.git
cd easy-search
# Install dependencies with uv
uv syncOr with pip:
python -m venv .venv
source .venv/bin/activate
pip install streamlit ddgs duckduckgo-searchuv run streamlit run app.pyThen open http://localhost:8501 in your browser.
The sidebar has five collapsible sections — all visible at once:
| Section | Default | Purpose |
|---|---|---|
| ➕ Add Operators | Open | Click Add next to any operator to add it to the Builder |
| 🎯 Query Templates | Collapsed | One-click pre-built operator stacks |
| ⚙️ Search Filters | Collapsed | Date range, region, deduplicate toggle |
| 💾 Saved Queries | Collapsed | Save/load named operator stacks |
| 📖 Operator Reference | Collapsed | Description, example, and tip for every operator |
- Add operators from the sidebar ➕ section, or open 🎯 Templates and load a full stack
- Edit values — fill in the input field on each operator card
- Reorder — use ↑ ↓ arrows to change operator positions
- Set filters — open ⚙️ Search Filters to set date range, region, and deduplication
- Search — adjust the max results slider (1–30), then click Search
- Copy query — the Query Preview box has a built-in copy button in its top-right corner
- Export results as CSV or JSON with the download buttons above the result cards
— Both formats contain three columns/fields:title,url,snippet - Filter by domain — type a domain (e.g.
github.com) to narrow displayed results without re-searching
- Build Query A and Query B independently in two columns
- Each side has its own template loader, operator picker, full builder, and results with export
- Enter a base query (e.g.
sentiment analysis dataset) - Select one or more target sites from the multiselect
- Set Results per site (1–10) with the slider
- Click Run Batch Search — searches run concurrently (up to 5 in parallel) and results are grouped by site
- Export each site's results individually as CSV or JSON
- Every search from the Builder tab is automatically saved (last 50)
- Click Reload to restore any past query's blocks back into the Builder
- Click 🗑️ to delete individual entries, or Clear history to wipe all
easy-search/
├── app.py # Streamlit UI — all features
├── pyproject.toml # Project metadata and dependencies
├── uv.lock # Locked dependency versions
└── README.md
| Package | Role |
|---|---|
| Streamlit | Web UI framework |
| ddgs | DuckDuckGo search API |
MIT — see LICENSE for details.