MCP Servers / development / vidlizer

vidlizer

JSON →
stdiononedevelopment

Extract structured JSON from video, images, and PDFs using local LLMs (Ollama, LM Studio, oMLX) or via OpenRouter. Runs fully offline

Install
How to run this server
[ { "cmd": "uvx (no", "imports": [] } ]
server path: vidlizer
Tools
What this server exposes
analyze_video
Analyzes a video file at the given path with optional parameters, returning an analysis ID and metadata.
list_analyses
Returns metadata for all stored analyses.
get_summary
Returns a text summary of an analysis at a specified level (brief, medium, or full).
get_step
Returns a single flow step from an analysis.
get_steps
Returns a range of steps from an analysis.
get_phase
Returns all steps in a named phase of an analysis.
search_analysis
Searches an analysis for steps matching a text query or keyword.
get_transcript
Returns a slice of the transcript from an analysis between specified start and end times.
get_full_analysis
Returns the full analysis including all flow steps and transcript.
delete_analysis
Deletes an analysis and returns a confirmation.
get_usage_stats
Returns per-model token and cost breakdown statistics.
clear_usage_stats
Resets the usage log and returns a confirmation.
Configuration
Environment & auth
authnone
envOPENAI_BASE_URL
envOPENROUTER_API_KEY
envOLLAMA_HOST
envOPENAI_API_KEY
envFALLBACK_BASE_URL
envFALLBACK_API_KEY
Resources
vidlizer — MCP Server · libregistry