Can vector search and metadata filters run together?

Yes. VideoVector supports search patterns that combine semantic retrieval with exact filters, nested field paths, run IDs, and index scope.

What is the difference between SQL search and vector search?

Vector search finds semantically relevant media moments. SQL search is for structured analysis, aggregations, selected columns, and repeatable query workflows over completed prompt-run data.

Technical solution

Video vector embedding search with metadata filters

Name: VideoVector
Brand: VectorMethods

VideoVector combines embedding search with exact structured metadata filters, SQL-style analysis, and agentic retrieval. Teams can search by meaning, visual reference, schema field, prompt run, index, or conversation scope.

On this pageThe search problem teams recognize The proposed retrieval workflow Search modes Technical indicators Why this is better than separate search tools Implementation path What this enables

Try now Read search workflows

The search problem teams recognize

Teams rarely struggle because they have no search box. They struggle because the search box only sees part of the media. It may see a transcript but not the image. It may see a manual tag but not the structured output. It may retrieve a broad asset but not the exact moment a reviewer needs.

VideoVector addresses that gap by combining embedding search with metadata filters, prompt-run scope, SQL-style exploration, and agentic retrieval. The goal is to let people ask business questions and still receive grounded, reviewable media results.

This is useful for executives who need faster answers from large media holdings, and for engineering managers who need one retrieval foundation that can serve product, operations, analyst, and automation workflows.

The proposed retrieval workflow

VideoVector supports a path from broad semantic discovery to precise, defensible result sets.

Search by meaning or example

Start with a natural-language query, image reference, or multimodal request to find relevant media moments without knowing the exact legacy tag or filename.

Constrain by business context

Narrow results by index, prompt run, nested metadata field, timestamp, or selected search fields so teams can control scope and precision.

Explore and hand off

Use SQL search for structured analysis, chat sessions for follow-up questions, and exports or APIs when results need to move into another workflow.

Search modes

Hybrid search lets operators combine semantic recall with structured precision.

Vector and multimodal search

Search by text, image, or fused multimodal query across indexed media context, embeddings, and prompt-run outputs.

Structured metadata filters

Constrain results with nested schema field paths such as scene.people[].emotion, run IDs, index IDs, and field-specific values.

SQL and agentic media search

Use SQL for analyst-style querying and chat sessions for scoped multi-turn retrieval over the same searchable media substrate.

Technical indicators

Use video vector embedding search when retrieval has to work across speech, visuals, schema outputs, and metadata_text.
Use hybrid vector and metadata search when production reviewers need both high recall and exact constraints.
Use agentic media search when users need follow-up questions, scoped turn history, streaming answers, and tool trace visibility.

Why this is better than separate search tools

Separate systems for transcript search, vector search, visual lookup, SQL exploration, and analyst assistants create inconsistent answers and duplicate engineering work.

VideoVector keeps those retrieval patterns connected to the same indexed media context and prompt-run outputs. Teams can use the simplest search mode for the job without losing access to the richer context underneath.

The result is a more practical path to production: one media foundation, multiple retrieval experiences, and clearer controls over scope, fields, filters, and handoff.

Implementation path

Start with the Search model page to choose direct, multimodal, SQL, filter, multi-run, or agentic retrieval.
Use schema-aware extraction and video embeddings first when search quality depends on domain-specific fields and generated media context.
Track query scopes deliberately by index ID, run ID, and field paths so results stay grounded in the intended corpus.

What this enables

Review teams can move from broad query to specific segment without manually rewatching long assets.
Analysts can ask follow-up questions while staying scoped to approved indexes, prompt runs, and result sets.
Engineering teams can build search experiences that combine semantic relevance with structured business rules.