dokpod

An AI knowledge vault that indexes UI screenshots, specs, and videos via MCP to stop token burn in Cursor and Claude

Community: Submitted by a user or imported; check the owner before granting accessOnlineAPI key requiredGlobalFreeRead-only

What it can do

  • List User Documents: List all documents in the user's Dokpod vault. Returns titles, tags, chunk counts, and status.
  • Semantic Search: Search the user's document vault semantically. Returns only the most relevant chunks (minimal tokens).
  • Find Documents By Name: Find vault files by exact or partial filename. Use when the user mentions a specific file name, PDF title, or screenshot name.

What data it sees

Do you need an account

An API key from the service settings is required

An AI knowledge vault that indexes UI screenshots, specs, and videos via MCP to stop token burn in Cursor and Claude

Server tool list (10)

Raw names from tools/list. Only developers need these.

list_user_documentsList all documents in the user's Dokpod vault. Returns titles, tags, chunk counts, and status.
semantic_searchSearch the user's document vault semantically. Returns only the most relevant chunks (minimal tokens).
find_documents_by_nameFind vault files by exact or partial filename. Use when the user mentions a specific file name, PDF title, or screenshot name.
get_chunkRetrieve a specific chunk by document_id and chunk_id.
summarize_documentGet the summary and metadata for a document without loading all chunks.
compare_documentsCompare two documents using summaries and key chunk excerpts.
list_mediaList images and videos in the user's vault with summaries and metadata.
semantic_search_mediaSearch images and videos semantically. Returns captions, frame descriptions, or transcript segments — minimal tokens.
get_media_descriptionGet full AI-generated description for an image or video (captions, frames, transcript).
get_relevant_framesFor a video, return only the frame descriptions most relevant to a query (low token usage).