Dokpod

A document indexing and retrieval tool that solves the problem of re-uploading files into AI coding editors. It indexes documents once and uses MCP search to return only relevant sections, reducing token usage by up to 97%. It targets developers and engineering teams using AI editors like Cursor and Claude Code, with plans for individuals and teams. It is delivered as a SaaS with a desktop option for local processing, and it differentiates by integrating via MCP and supporting multiple file types including images and videos.

Key features

• Upload PDFs, Word docs, spreadsheets, images, videos
• Import from URLs, YouTube, Google Drive, GitHub
• Visual descriptions and text detection for images
• Indexes video transcripts and scenes
• Semantic search returns relevant sections only
• MCP integration for Cursor and Claude Code
• Private vault with team sharing option
• Token savings of 80-97%