35 lines
1.2 KiB
Markdown
35 lines
1.2 KiB
Markdown
# Research Brief LangGraph Agent
|
||
|
||
## What it does
|
||
This repository contains a **LangGraph** agent that builds a short research brief for any given topic. The brief is produced in three stages:
|
||
|
||
1. **Outline** – the LLM generates 4‑5 concise bullet points.
|
||
2. **Research step** – each outline item triggers one web search via Tavily and produces a short note (5–8 sentences).
|
||
3. **Synthesize** – all notes are combined into a coherent brief with headings.
|
||
|
||
The agent is fully automated; running the script will print the outline, individual notes and the final brief.
|
||
|
||
## File structure
|
||
- `agent.py` – main implementation of the LangGraph graph.
|
||
- `requirements.txt` – Python dependencies.
|
||
- `README.md` – this documentation.
|
||
|
||
## Installation
|
||
```bash
|
||
pip install -r requirements.txt
|
||
```
|
||
|
||
Create a `.env` file in the project root with your Tavily API key:
|
||
```
|
||
TAVILY_API_KEY=your_key_here
|
||
JOURNAL_MCP_PAT=your_brojs_pat # required for LLM access
|
||
```
|
||
|
||
## Usage
|
||
```bash
|
||
python agent.py
|
||
```
|
||
The script will run a demo with the default topic *"Как студенту безопасно подключать MCP к LangChain"*.
|
||
|
||
You can also import `BriefGraph` from `agent` and use it in your own code.
|