# langchain-parallel-web
LangChain tool and toolkit for Parallel Web search. It exposes a tool that performs web research and returns ranked, compressed excerpts from multiple sources, plus a toolkit to easily wire it into agents.
## Installation
```bash
pip install -U langchain-parallel-web
```
Set your Parallel API key:
```bash
export PARALLEL_API_KEY="your-api-key"
```
You can also pass the key directly to the classes shown below.
## Tool: Parallel Web Search
`ParallelWebTool` exposes the Parallel Web search as a LangChain tool.
```python
from langchain_parallel_web.tools import ParallelWebTool
# Option 1: rely on PARALLEL_API_KEY in env
tool = ParallelWebTool()
# Option 2: pass the key explicitly
# tool = ParallelWebTool(parallel_api_key="your-api-key")
result_json = tool.invoke({
"objective": "Find when the UN was founded; prefer official sources.",
"search_queries": [
"United Nations founding year",
"When was the UN founded official site"
],
})
print(result_json)
```
### Configuration
You can control search behavior via `ParallelWebSearchConfig`.
```python
from langchain_parallel_web.parallel_web_wrapper import ParallelWebSearchConfig
from langchain_parallel_web.tools import ParallelWebTool
config = ParallelWebSearchConfig(
max_results=10, # up to 40
processor="base", # or "pro"
max_chars_per_result=1500
)
tool = ParallelWebTool(config=config)
```
## Toolkit: Agent Integration
`ParallelWebToolkit` returns a ready-to-use list of tools for agent setups.
```python
from langchain_parallel_web.toolkits import ParallelWebToolkit
# Pass the key explicitly for the toolkit
toolkit = ParallelWebToolkit(parallel_api_key="your-api-key")
tools = toolkit.get_tools()
# Use `tools` with your LangChain agent or runnable graph
```
## Direct API Wrapper
For lower-level control, call the wrapper directly.
```python
from langchain_parallel_web.parallel_web_wrapper import (
ParallelWebAPIWrapper,
ParallelWebSearchInput,
)
wrapper = ParallelWebAPIWrapper() # uses PARALLEL_API_KEY from env
# or: wrapper = ParallelWebAPIWrapper(parallel_api_key="your-api-key")
payload = ParallelWebSearchInput(
objective="Summarize key announcements from OpenAI DevDay 2023.",
search_queries=["OpenAI DevDay 2023 announcements", "OpenAI DevDay keynote summary"],
)
out = wrapper.search(payload)
print(out.model_dump())
```
Async usage is supported:
```python
import asyncio
from langchain_parallel_web.parallel_web_wrapper import (
ParallelWebAPIWrapper,
ParallelWebSearchInput,
)
async def main():
wrapper = ParallelWebAPIWrapper()
out = await wrapper.asearch(
ParallelWebSearchInput(
objective="Latest SEC filing summary for NVDA.",
search_queries=["NVIDIA latest 10-Q", "NVDA SEC filing"],
)
)
print(out.model_dump())
asyncio.run(main())
```
Raw data
{
"_id": null,
"home_page": "https://github.com/sanjeevhalyal/langchain-parallel-web",
"name": "langchain-parallel-web",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.9",
"maintainer_email": null,
"keywords": "langchain, parallel, web, llm, ai",
"author": "Sanjeev Halyal",
"author_email": "1448731+sanjeevhalyal@users.noreply.github.com",
"download_url": "https://files.pythonhosted.org/packages/9b/32/118883fbe8458516e5c5514c5033e07782f82d5f14c09b8d67d4e25cde9e/langchain_parallel_web-0.0.1a0.tar.gz",
"platform": null,
"description": "# langchain-parallel-web\n\nLangChain tool and toolkit for Parallel Web search. It exposes a tool that performs web research and returns ranked, compressed excerpts from multiple sources, plus a toolkit to easily wire it into agents.\n\n## Installation\n\n```bash\npip install -U langchain-parallel-web\n```\n\nSet your Parallel API key:\n\n```bash\nexport PARALLEL_API_KEY=\"your-api-key\"\n```\n\nYou can also pass the key directly to the classes shown below.\n\n## Tool: Parallel Web Search\n\n`ParallelWebTool` exposes the Parallel Web search as a LangChain tool.\n\n```python\nfrom langchain_parallel_web.tools import ParallelWebTool\n\n# Option 1: rely on PARALLEL_API_KEY in env\ntool = ParallelWebTool()\n\n# Option 2: pass the key explicitly\n# tool = ParallelWebTool(parallel_api_key=\"your-api-key\")\n\nresult_json = tool.invoke({\n \"objective\": \"Find when the UN was founded; prefer official sources.\",\n \"search_queries\": [\n \"United Nations founding year\",\n \"When was the UN founded official site\"\n ],\n})\nprint(result_json)\n```\n\n### Configuration\nYou can control search behavior via `ParallelWebSearchConfig`.\n\n```python\nfrom langchain_parallel_web.parallel_web_wrapper import ParallelWebSearchConfig\nfrom langchain_parallel_web.tools import ParallelWebTool\n\nconfig = ParallelWebSearchConfig(\n max_results=10, # up to 40\n processor=\"base\", # or \"pro\"\n max_chars_per_result=1500\n)\n\ntool = ParallelWebTool(config=config)\n```\n\n## Toolkit: Agent Integration\n\n`ParallelWebToolkit` returns a ready-to-use list of tools for agent setups.\n\n```python\nfrom langchain_parallel_web.toolkits import ParallelWebToolkit\n\n# Pass the key explicitly for the toolkit\ntoolkit = ParallelWebToolkit(parallel_api_key=\"your-api-key\")\n\ntools = toolkit.get_tools()\n\n# Use `tools` with your LangChain agent or runnable graph\n```\n\n## Direct API Wrapper\n\nFor lower-level control, call the wrapper directly.\n\n```python\nfrom langchain_parallel_web.parallel_web_wrapper import (\n ParallelWebAPIWrapper,\n ParallelWebSearchInput,\n)\n\nwrapper = ParallelWebAPIWrapper() # uses PARALLEL_API_KEY from env\n# or: wrapper = ParallelWebAPIWrapper(parallel_api_key=\"your-api-key\")\n\npayload = ParallelWebSearchInput(\n objective=\"Summarize key announcements from OpenAI DevDay 2023.\",\n search_queries=[\"OpenAI DevDay 2023 announcements\", \"OpenAI DevDay keynote summary\"],\n)\n\nout = wrapper.search(payload)\nprint(out.model_dump())\n```\n\nAsync usage is supported:\n\n```python\nimport asyncio\nfrom langchain_parallel_web.parallel_web_wrapper import (\n ParallelWebAPIWrapper,\n ParallelWebSearchInput,\n)\n\nasync def main():\n wrapper = ParallelWebAPIWrapper()\n out = await wrapper.asearch(\n ParallelWebSearchInput(\n objective=\"Latest SEC filing summary for NVDA.\",\n search_queries=[\"NVIDIA latest 10-Q\", \"NVDA SEC filing\"],\n )\n )\n print(out.model_dump())\n\nasyncio.run(main())\n```\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "ParallelWeb integration for LangChain",
"version": "0.0.1a0",
"project_urls": {
"Homepage": "https://github.com/sanjeevhalyal/langchain-parallel-web",
"Issues": "https://github.com/sanjeevhalyal/langchain-parallel-web/issues",
"Release Notes": "https://github.com/sanjeevhalyal/langchain-parallel-web/releases",
"Repository": "https://github.com/sanjeevhalyal/langchain-parallel-web",
"Source Code": "https://github.com/sanjeevhalyal/langchain-parallel-web"
},
"split_keywords": [
"langchain",
" parallel",
" web",
" llm",
" ai"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "539a8a53592602faf2c3f469f71250d28da2fa4095cc714ab27af926133dfa48",
"md5": "ba8a2abdb115f493997ec086d814c759",
"sha256": "0f813b535116d8b4517fc65ee9bba47121cd5a89a1658093f2278e768584df62"
},
"downloads": -1,
"filename": "langchain_parallel_web-0.0.1a0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "ba8a2abdb115f493997ec086d814c759",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.9",
"size": 10661,
"upload_time": "2025-08-24T05:19:17",
"upload_time_iso_8601": "2025-08-24T05:19:17.036593Z",
"url": "https://files.pythonhosted.org/packages/53/9a/8a53592602faf2c3f469f71250d28da2fa4095cc714ab27af926133dfa48/langchain_parallel_web-0.0.1a0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "9b32118883fbe8458516e5c5514c5033e07782f82d5f14c09b8d67d4e25cde9e",
"md5": "82249a047e8edc79c08e2f42ed3cdbf7",
"sha256": "d47f856862877005fca7acf40041bd4b696ecd7c5535f5c276a10df2b61b0d5f"
},
"downloads": -1,
"filename": "langchain_parallel_web-0.0.1a0.tar.gz",
"has_sig": false,
"md5_digest": "82249a047e8edc79c08e2f42ed3cdbf7",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.9",
"size": 9232,
"upload_time": "2025-08-24T05:19:18",
"upload_time_iso_8601": "2025-08-24T05:19:18.555175Z",
"url": "https://files.pythonhosted.org/packages/9b/32/118883fbe8458516e5c5514c5033e07782f82d5f14c09b8d67d4e25cde9e/langchain_parallel_web-0.0.1a0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-08-24 05:19:18",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "sanjeevhalyal",
"github_project": "langchain-parallel-web",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "langchain-parallel-web"
}
JSON
