AI Source Scraper 作者: Janna Joceli Omena
AI Source Scrape captures, structures, and compares every cited web source — with metadata — from AI search interfaces, namely Claude, Gemini, and ChatGPT.
擴充套件後設資料
畫面擷圖
關於此擴充套件
AI Source Scraper
A research tool for capturing and comparing sources surfaced by AI search interfaces.
Designed for digital methods and AI search research, the extension turns source and activity panels from supported AI platforms into structured datasets while preserving information about the conditions under which each response was generated.
Rather than collecting links alone, AI Source Scraper lets researchers annotate each capture with methodological metadata:
- Capture mode: Longitudinal or Comparative
- Search condition: AI search or Agentic AI search
- Prompt condition: Isolated, Initial, or Follow-up
- Prompt framing: Underspecified, Program, Anti-program, or Ambiguous
- AI fieldwork observations: researcher notes recorded during data collection
Captures accumulate within a research session and can be exported as CSV, JSON, and TXT, allowing researchers to merge, filter, compare, and analyse source-selection patterns across prompts, platforms, search conditions, and moments in time.
This research tool operationalises digital methods such as the [Longitudinal Prompt Method for Understanding AI Search] (https://www.researchgate.net/publication/408580170_Longitudinal_Prompt_Method_for_Understanding_AI_Search), supporting projects that seek to advance source studies of single or cross-AI models by turning the citation layer of AI chatbots into structured data.
Two research designs
Longitudinal captures are designed to investigate how AI responses, sources, or interfaces develop and change over time or across conversational sequences.
Comparative captures support research designs in which conditions are deliberately varied, for example by comparing AI search with Agentic AI search or testing different prompt framings.
The same metadata structure is available in both modes, making it possible to build comparable datasets while keeping the methodological conditions of each capture explicit.
Supported research contexts
AI Source Scraper is intended for research involving AI-mediated web search, source selection, citation practices, conversational search, and the changing relationship between prompts, search conditions, and retrieved web sources.
ChatGPT and Gemini can be studied in logged-in or logged-out conditions where their interfaces permit anonymous use. Claude.ai requires an account to conduct conversations.
Data and privacy
The extension processes captured research data locally in the browser. Researchers decide when to export or clear their session data.
Citation
Omena, J. J. (2026). AI Source Scraper [Computer software]. Zenodo.
https://doi.org/10.5281/zenodo.20945556
Source code:
https://github.com/jannajoceli/ai-source-scraper
A research tool for capturing and comparing sources surfaced by AI search interfaces.
Designed for digital methods and AI search research, the extension turns source and activity panels from supported AI platforms into structured datasets while preserving information about the conditions under which each response was generated.
Rather than collecting links alone, AI Source Scraper lets researchers annotate each capture with methodological metadata:
- Capture mode: Longitudinal or Comparative
- Search condition: AI search or Agentic AI search
- Prompt condition: Isolated, Initial, or Follow-up
- Prompt framing: Underspecified, Program, Anti-program, or Ambiguous
- AI fieldwork observations: researcher notes recorded during data collection
Captures accumulate within a research session and can be exported as CSV, JSON, and TXT, allowing researchers to merge, filter, compare, and analyse source-selection patterns across prompts, platforms, search conditions, and moments in time.
This research tool operationalises digital methods such as the [Longitudinal Prompt Method for Understanding AI Search] (https://www.researchgate.net/publication/408580170_Longitudinal_Prompt_Method_for_Understanding_AI_Search), supporting projects that seek to advance source studies of single or cross-AI models by turning the citation layer of AI chatbots into structured data.
Two research designs
Longitudinal captures are designed to investigate how AI responses, sources, or interfaces develop and change over time or across conversational sequences.
Comparative captures support research designs in which conditions are deliberately varied, for example by comparing AI search with Agentic AI search or testing different prompt framings.
The same metadata structure is available in both modes, making it possible to build comparable datasets while keeping the methodological conditions of each capture explicit.
Supported research contexts
AI Source Scraper is intended for research involving AI-mediated web search, source selection, citation practices, conversational search, and the changing relationship between prompts, search conditions, and retrieved web sources.
ChatGPT and Gemini can be studied in logged-in or logged-out conditions where their interfaces permit anonymous use. Claude.ai requires an account to conduct conversations.
Data and privacy
The extension processes captured research data locally in the browser. Researchers decide when to export or clear their session data.
Citation
Omena, J. J. (2026). AI Source Scraper [Computer software]. Zenodo.
https://doi.org/10.5281/zenodo.20945556
Source code:
https://github.com/jannajoceli/ai-source-scraper
由 1 位評論者給出 0 分
權限與資料
必要權限：
- 存取您在 claude.ai 的資料
- 存取您在 gemini.google.com 的資料
- 存取您在 chatgpt.com 的資料
- 存取您在 chat.openai.com 的資料
收集下列資料：
- 開發者聲稱此擴充套件不要求收集任何資料。
更多資訊