OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Search & Research AI

Chronos — Alibaba-NLP

CHRONOS is an open-source Python research implementation that retrieves news through iterative question generation and produces chronological timeline summaries. Its repository also provides the Open-TLS dataset and scripts for running experiments.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Chronos — Alibaba-NLP — OpenViST project profile in Search & Research AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use CHRONOS when a topic needs a chronological account assembled from retrieved news, or when developing and evaluating open-domain timeline-summarization workflows.

Who it is for

  • Researchers studying open-domain timeline summarization
  • Analysts and developers prototyping topic-based news timelines
  • Teams evaluating timeline generation with the Open-TLS dataset

Documented uses

  • Retrieve news relevant to a topic through successive questions about the topic and retrieved documents.
  • Arrange retrieved reporting into a chronological summary.
  • Run timeline-summarization experiments using Open-TLS and inspect retrieved news, generated timelines, and evaluation scores.
Before you choose

Requirements & limits

The repository describes an open-source implementation. Its documented workflow requires external API keys for a Qwen or GPT model and Bing Web Search; a Jina key is needed for the optional full-page-reading path.

  • The documented workflow depends on both a model API and Bing Web Search.
  • The question-example pool can be generated for datasets in the data directory, or users can use the examples supplied for Crisis, T17, and Open-TLS.
  • The repository identifies Jina as an optional key for reading full pages rather than only search snippets.
  • Replacing Qwen or GPT with another model backend requires modification; the repository does not demonstrate that substitution.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Run CHRONOS on the Open-TLS dataset

Configure the documented Python workflow and run a timeline-summarization experiment on Open-TLS.

Before you start

  • The CHRONOS repository and Python environment
  • Dependencies installed from requirements.txt
  • An API key for Qwen or GPT and a Bing Web Search API key
  • The Open-TLS question examples, either generated for the dataset or taken from data/question_examples.json

Steps

  1. Install dependencies with pip install -r requirements.txt.

  2. Generate a topic-question example pool with python question_exampler.py, or use the provided data/question_examples.json.

  3. Add the model-provider key in src/model.py and the Bing Web Search key in src/searcher.py.

  4. Run main.py with the Open-TLS dataset, a model name, a maximum self-questioning round, an output directory, and the question-examples option.

Expected outcome

The configured output directory contains retrieved news, generated timelines, and evaluation scores.

Guide limitations

  • The documented model choices are Qwen and GPT, and news retrieval uses Bing Web Search.
  • A Jina key is described as necessary when using full-page content instead of snippets.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
Alibaba-NLP/CHRONOS · Official homepage title · retained in its source language
How we review sourcesSuggest a correction