Metadata-Version: 2.5
Name: langchain-quanticdata
Version: 0.1.0
Summary: LangChain integration for QuanticData — document loaders, a web retriever and an agent toolkit for scraping, SERP search, crawling and 74 ready-made collectors through residential proxies.
Project-URL: Homepage, https://quanticdata.io
Project-URL: Documentation, https://quanticdata.io/docs/
Project-URL: Repository, https://github.com/quantumproxies/langchain-quanticdata
Author-email: QuanticData <support@quanticdata.io>
License: MIT
License-File: LICENSE
Keywords: agents,document-loader,langchain,quanticdata,retriever,serp,web-scraping
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Internet :: WWW/HTTP
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Python: >=3.9
Requires-Dist: langchain-core>=0.3
Requires-Dist: quanticdata>=0.1.0
Description-Content-Type: text/markdown

# langchain-quanticdata

LangChain integration for [QuanticData](https://quanticdata.io): give an agent
live web data — pages as clean Markdown, structured SERP results, whole-site
crawls, SEO audits, 74 ready-made collectors and prompt-driven datasets — all
routed through residential proxies with real-browser TLS fingerprints, billed
only on success.

```bash
pip install langchain-quanticdata
```

Set your key once (free monthly usage, no card required — get one at
[quanticdata.io](https://quanticdata.io)):

```bash
export QUANTICDATA_API_KEY=qd_live_your_key_here
```

## Document loaders

```python
from langchain_quanticdata import QuanticDataLoader, QuanticDataCrawlLoader

# one or more URLs, rendered only when the page needs it
docs = QuanticDataLoader(["https://example.com"], content_mode="smart").load()
print(docs[0].page_content)          # Markdown
print(docs[0].metadata["title"], docs[0].metadata["engine"])

# a whole site, breadth-first
docs = QuanticDataCrawlLoader("https://docs.example.com", limit=50, depth=2).load()
```

## Retriever

Search the live web and get the top sources back as documents — useful when an
index would go stale:

```python
from langchain_quanticdata import QuanticDataRetriever

retriever = QuanticDataRetriever(k=3, country="us")
docs = retriever.invoke("what changed in the EU AI act")
```

## Agent toolkit

```python
from langchain.agents import create_agent
from langchain_quanticdata import QUANTICDATA_TOOLS

agent = create_agent(model, tools=QUANTICDATA_TOOLS)
agent.invoke({"messages": [("user", "Find the 5 cheapest Nintendo Switch OLED offers in the US")]})
```

| Tool | What it does |
|------|--------------|
| `quanticdata_scrape` | A page as clean Markdown; `render=True` for JS-heavy sites |
| `quanticdata_extract` | Structured JSON from a page, driven by a plain-language instruction |
| `quanticdata_search` | Google results with title, link and snippet |
| `quanticdata_research` | Search plus the top pages, assembled into citation-ready context |
| `quanticdata_map_site` | Every URL of a site from its sitemaps, filterable |
| `quanticdata_seo_audit` | A URL with and without JavaScript, and the difference |
| `quanticdata_list_collectors` | The catalogue of ready-made collectors |
| `quanticdata_run_collector` | Run one by slug — Amazon, Google Maps, LinkedIn jobs, app stores… |
| `quanticdata_create_dataset` | A validated dataset from a description of the rows you want |

Import any of them individually if you want a smaller surface:

```python
from langchain_quanticdata import quanticdata_search, quanticdata_scrape
```

## Notes

- Failed calls (blocked pages, captchas) are never billed.
- The tools call the public QuanticData API with your own key; nothing else is
  stored or proxied.
- Full parameter reference: [quanticdata.io/docs](https://quanticdata.io/docs/).

MIT licensed.
