A Python library and runnable Jupyter textbook implementing 35 agentic AI architectures, from Reflexion and ReAct to MemGPT and BrowserAgent, on LangGraph.
rotemweiss57/gpt-newspaper
An AI newsroom of LLM agents that researches topics you pick, writes cited articles, and prints a personalized newspaper front page.
- Stars
- 1,486
- Forks
- 205
- Open issues
- 22
- Last commit
- 7 days ago
What is it?
GPT Newspaper generates a personalized newspaper from a list of topics. Reporter agents search the news through Tavily, a curator filters off-topic results, writers draft articles from full source text, and a critic sends drafts back for revision before an editor-in-chief assembles the front page. Inline citations link factual claims to numbered sources.
The pipeline is built on LangGraph and uses LangChain's model interface, so OpenAI, Anthropic, Google, Groq, Mistral, xAI, DeepSeek and local Ollama models can be selected with a single provider:model string. An art director agent can commission illustrations in ink, engraving or risograph styles, or use the sources' news photos instead.
The project offers a web app with live progress streaming, a streaming HTTP API, a command-line interface, and LangGraph Studio support. Editions can be printed or saved as PDF and are stored in an archive.
Key features
- 01Plans news searches per topic and runs them in parallel through Tavily.
- 02Filters and ranks sources, widening the time range when a topic has little recent news.
- 03Drafts articles with inline citations and a critic loop for revisions.
- 04Supports many LLM providers through one model string, including local Ollama.
- 05Generates illustrations in several styles or uses source photos.
- 06Streams progress via a web UI, API, or CLI.
Use cases
- Receiving a daily personal newspaper on topics you follow, generated by cron via the CLI.
- Prototyping a multi-agent research and writing pipeline on LangGraph.
- Producing cited news digests in languages other than English.
Quick start
- 1
Clone the repository
git clone https://github.com/rotemweiss57/gpt-newspaper.git - 2
Change into the project directory
cd gpt-newspaper - 3
Install the Python dependencies
pip install -r requirements.txt - 4
Start the web app, then open http://localhost:8000 and add your keys
python app.py
Things to know
- Requires a Tavily API key, and a model provider key or a local Ollama install.
- Illustrations need an OpenAI key even when another provider writes the articles.
- Articles are AI-generated and can contain mistakes, so sources must be checked.
- Described as an experimental project provided as is.
In collections
Topics
FAQ
- What is gpt-newspaper?
- An AI newsroom of LLM agents that researches topics you pick, writes cited articles, and prints a personalized newspaper front page. GPT Newspaper generates a personalized newspaper from a list of topics. Reporter agents search the news through Tavily, a curator filters off-topic results, writers draft articles from full source text, and a critic sends drafts back for revision before an editor-in-chief assembles the front page. Inline citations link factual claims to numbered sources.
- Is gpt-newspaper free to use? What license does it have?
- gpt-newspaper is open source under the MIT license. Check the license terms before using it in commercial products.
- Is gpt-newspaper still maintained?
- Actively maintained. The last commit to gpt-newspaper was on Nov 1, 2026.
- What can I use gpt-newspaper for?
- Receiving a daily personal newspaper on topics you follow, generated by cron via the CLI. Prototyping a multi-agent research and writing pipeline on LangGraph. Producing cited news digests in languages other than English.
- How do I install gpt-newspaper?
- Clone the repository: git clone https://github.com/rotemweiss57/gpt-newspaper.git Change into the project directory: cd gpt-newspaper Install the Python dependencies: pip install -r requirements.txt Start the web app, then open http://localhost:8000 and add your keys: python app.py
- Where can I get gpt-newspaper?
- The source code and installation instructions are on GitHub: https://github.com/rotemweiss57/gpt-newspaper
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A RAG tool that summarizes and answers questions across mixed documents (PDFs, audio, YouTube, Anki, Logseq, web pages) using any LLM provider.
An opinionated, batteries-included agent harness for Python that runs long multi-step tasks out of the box on LangGraph.
Multi-agent research assistant built on LangGraph that generates hypotheses, runs data analysis and visualization, and writes research reports.
FastAPI backend for a conversational RAG agent that embeds documents into Qdrant, runs semantic and hybrid search, and answers questions with LLMs.
Open-source ChatBI tool that turns natural-language questions into SQL queries, charts and data insights, built on LangChain/LangGraph with MCP-based multi-agent workflows.