Extract names with their locations on a page
Provide a public page or authorized copy and define the entity types to extract. Produce a list with section locations.
Execution environment and usage
AI / software
ChatGPT text analysis + human text-editor save
Where to operate
ChatGPT text analysis, followed by a human save in a local text editor.
Getting started
- Open the public page or authorized copy in your browser and copy the requested text with headings, sections and URL.
- Start a chat at chatgpt.com, paste the source and original prompt, and specify entity types, scope and output format.
- Check names and locations against the source. Copy the checked result into a local text editor and use Save As to save a UTF-8 .txt at the chosen location.
- Reopen the file to verify it. This route assigns page collection and saving to you; it does not claim agent access to your disk.
Setup and plan requirements
Specified entity scope and readable source; writable output path only if a real local file is requested.
Prompt portability limits
Preserves the extracted list and local file, with page collection and saving performed by the user. It removes automated retrieval/direct file writing.
Usage limitations
API client toolset exists; distinguish from separately configured Cowork computer use. Historical tool/source attribution is retained. Not hands-on tested; verify plan, region, workspace and site access separately.
Environment sources
適用情境 / When to use
Provide a public page or authorized copy and define the entity types to extract. Produce a list with section locations. Supplied text is enough for a draft; saving a local file requires an authorized file tool. Source record: williamtkelley reports using Claude computer use to extract names from a Wikipedia page into a text file.
步驟 / Steps
- Prepare a public page or authorized copy, extraction scope and writable test location.
- Extract the specified entity type with sections, original spellings and occurrence locations.
- Check source passages and the list, separating namesakes, aliases and uncertain entries.
提示詞 / Prompt
Extract [entity type and scope] from [public page or authorized copy], retaining names, locations and ambiguity notes. Create a local draft in [format]. Add no external personal information, guess no identities and do not merge similar spellings automatically; disclose uncovered page regions.
留意事項 / Cautions
- A public name does not authorize collecting private contact details.
- Do not add desktop permissions when text extraction suffices.
來源限制 / Evidence limits
The author says it seemed to work without completeness validation; historical charges do not establish present costs.
流程說明 / Workflow note
These steps and prompts are original, untested exercises by this site, not the author’s verbatim workflow. Reported results do not guarantee exercise outcomes.
原始工具組合 / Original tool stack
Claude computer use; Firefox; local Docker demonstration
Source checked: 2026-10-19