📚 OKF Wiki Builder 🗂 قضية سرقة عقد بيسان

🗂 قضية سرقة عقد بيسان

💬 Chat 🖨 Export HTML

ملف قضية جنائية كامل — من بلاغ السرقة حتى الحكم. قضية افتراضية بالكامل أُعدّت لعرض إمكانات النظام؛ جميع الأسماء والوقائع من نسج الخيال.

Workspace demo-beisan-case · Model gpt-4o-mini · Bundle at /home/runner/workspace/data/workspaces/demo-beisan-case/wiki · Postgres mirror: disabled

166
Entities
383
Relationships
24
Relationship types
12
Sources ingested

Knowledge by type

Event · 35 Action · 30 Date · 22 Place · 19 Organization · 18 Document · 16 Person · 15 Artifact · 10 Technology · 1

Ingest sources

Add documents from this computer. They are copied into the workspace, so the originals are never touched.

Which formats can I add?
📝 .md .txtread directly
📕 .pdftext layer read; scanned or garbled-Arabic PDFs go to OCR automatically
📘 .docxparagraphs and tables
✉️ .msgOutlook email — subject, sender, date, body, attachment names
🖼️ .png .jpg .tif .tiffread with Mistral OCR (multi-page TIFF page by page)
Runs AI over each source · costs tokens

Point at a folder and every supported file inside it (including subfolders) is ingested.

💡 This path is read on the machine running the app, not on your own computer — they're the same when you run it locally. The folder is remembered, so “Reprocess changed files” can pick up later edits.

Runs AI over each source · costs tokens

Upload an archive and it's extracted into the workspace, keeping its folder structure.

💡 Handy for moving a document set between machines. Entries that would escape the workspace directory are rejected.

Runs AI over each source · costs tokens

Fetch a web page and keep only its main content — header, menus, sidebars, footers, cookie banners and comments are stripped away.

💡 Nothing is added to the wiki yet: you get the cleaned Markdown to read and edit first, then choose whether to process it.

Fetching costs nothing — AI runs only after you approve.

Paste notes, minutes or an email thread. It's saved as a normal Markdown source, so it's cited and reprocessable like any file.

💡 Give it a descriptive name — it becomes the source filename every fact from this text is cited against.

Runs AI over each source · costs tokens
What happens next: scan extract text AI analysis write pages render bundle

Job status

State: idle 






  





  
  re-scans ingested folder(s), extracts only new/changed sources






Run history





  
StartedStatusModelLanguage
             FilesTokens in / outCostPrompts used

  
  
  

    2026-07-25 20:57
    done
        
    gpt-5.4-mini
    العربية
    12 (289E/423R)
    
        87,788 / 79,319
        
        
    n/a
    
      entity_extraction:def·59ba12a3 relationship_extraction:def·5ef74667 wiki_generation:def·3be0191e deduplication:def·81611a1b timeline_generation:def·273fa63b data_extraction:def·ebcebd7b data_relevance:def·4b22452c chat_answer:def·a07aabb5 chat_retrieval:def·f342af2d 
    
  

  
  








Recent wiki activity





Danger zone