Access and Synthesize Research Data with Linkup
Run configurable research queries across data sources with Linkup, tailored for LLM agents.
Why it matters
Leverage the Linkup Research API to gather and synthesize information from diverse data sources, providing relevant answers and sources for LLM agents.
Outcomes
What it gets done
Integrate with the Linkup Research API.
Configure research depth and output format (search results, sourced answers, or structured data).
Utilize the tool within an agent to query for specific information.
Retrieve and process research findings for LLM agent consumption.
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/li-tool-tools-linkup-research | bash Overview
Linkup Research Tool
A LlamaIndex tool for the Linkup research API, with configurable research depth and a choice of result, sourced-answer, or structured output. Use deep depth when completeness matters more than speed, and sourcedAnswer when citations are needed.
What it does
The Linkup Research Tool connects a LlamaIndex agent to Linkup, a research API purpose-built for LLM agents that integrates with diverse data sources to deliver a relevant research experience. LinkupToolSpec is configured with an API key plus two settings that shape how results come back: depth, choosing standard for a faster result or deep for a slower but more complete one, and output_type, choosing searchResults for a list of results relative to the query, sourcedAnswer for a synthesized answer plus its sources, or structured to get results shaped to a specific schema you supply via structured_output_schema.
The source's own worked example queries "which women were awarded the Physics Nobel Prize," showing the tool answering a specific factual research question through an agent rather than returning raw search results only.
When to use - and when NOT to
Use depth=standard when you need a faster answer and can accept a lighter research pass; use depth=deep when completeness matters more than speed. Choose output_type=sourcedAnswer when you want a direct answer with citations, searchResults when you want the raw list of matches, or structured when downstream code expects a specific schema. Do not use structured_output_schema unless output_type is set to structured - it is only relevant in that mode.
Capabilities
Runs research queries through Linkup with configurable depth (standard/deep) and output shape (search results list, a sourced answer with citations, or a schema-structured result).
How to install
pip install llama-index-tools-linkup-research
Requires an API key from the Linkup dashboard.
Who it's for
Developers building LlamaIndex agents that need configurable, citation-aware research answers - trading off speed for completeness, and choosing between raw results, a cited answer, or a structured schema.
Source README
Linkup Research Tool
Linkup is a robust research API tailored specifically for LLM Agents. It seamlessly integrates with diverse data sources to ensure a superior, relevant research experience.
- you need to obtain an API key on the Linkup dashboard
Quick Start:
pip install llama-index-tools-linkup-research
import os
from llama_index.core.agent.workflow import FunctionAgent
from llama_index.llms.openai import OpenAI
from llama_index.tools.linkup_research.base import LinkupToolSpec
### structured_schema=json.dumps(your schema here) # Only if output type is structured
### Initialisation of the tool
linkup_tool = LinkupToolSpec(
api_key="your Linkup API Key",
depth="", # Choose (standard) for a faster result (deep) for a slower but more complete result.
output_type="", # Choose (searchResults) for a list of results relative to your query, (sourcedAnswer) for an answer and a list of sources, or (structured) if you want a specific schema.
# structured_output_schema=structured_schema # Only if output type is structured
)
### Creation of the agent
agent = FunctionAgent(
tools=linkup_tool.to_tool_list(),
llm=OpenAI(model="gpt-4o"),
)
### Query for the agent
await agent.run(
"Can you tell me which women were awarded the Physics Nobel Prize"
)
This loader is designed to be used as a way to load data as a Tool in an Agent.
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.