video-research-mcp

Overview Schema Source Related Servers Discussions Admin

Video Analyze

video_analyze

Idempotent

Analyze YouTube videos or local video files with custom instructions to extract structured JSON insights, summaries, commands, recipes, or schema-defined results.

Instructions

Analyze a video (YouTube URL or local file) with any instruction.

Provide exactly one of url or file_path. Uses Gemini's structured output for reliable JSON responses. Pass a custom output_schema to control the response shape, or use the default VideoResult schema.

When strict_contract=True, runs the full contract pipeline: analysis with strict Pydantic models, parallel strategy/concept-map generation, artifact rendering, and structural quality gates. Factual/media review and observed coverage remain explicit pending/unknown fields. Returns richer output but takes longer.

Input Schema

Name	Required	Description	Default
`fps`	No	Local files only: requested static sampling frames per second; observed coverage remains unknown
`url`	No	YouTube video URL.
`dry_run`	No	Plan source reads/sends without provider calls
`file_path`	No	Path to a local video file.
`use_cache`	No	Use cached results
`end_offset`	No	Local source interval end using h/m/s units; timestamps retain the original source origin
`instruction`	No	What to analyze — e.g. 'summarize key points', 'extract all CLI commands shown', 'list all recipes and ingredients'	Provide a comprehensive analysis of this video.
`start_offset`	No	Local source interval start, using h/m/s units, for example '27m' or '1m30.5s'
`output_fields`	No	Select top-level response fields; complete source/citation/provenance carriers remain included
`output_schema`	No	Optional JSON Schema for the response. If omitted, uses default VideoResult schema.
`thinking_level`	No	Gemini thinking depth.	high
`strict_contract`	No	Enable strict contract pipeline with quality gates, artifact rendering, and structural validation. Factual/media review and observed coverage remain explicitly pending or unknown. Produces richer output with strategy report, concept map, and HTML/Markdown artifacts.
`execution_budget`	No	Explicit max_calls/max_tokens/max_output_tokens/max_frames/max_windows and start_ms/end_ms/fps. Bounded execution skips optional enrichment and cache reuse.
`transcript_limit`	No	Page a top-level transcript string without truncating stored evidence or citations
`transcript_offset`	No	Original custom-schema transcript string offset in Unicode code points

Output Schema

Name	Required	Description	Default
No arguments

Schema Changelog

Changes observed during successful MCP inspections.

Changed9 schema fields changedv0.7.1Oct 5, 2026

addedInput schema / properties / dry_run

Added value: +{
+  "default": false,
+  "description": "Plan source reads/sends without provider calls",
+  "type": "boolean"
+}

addedInput schema / properties / end_offset

Added value: +{
+  "anyOf": [
+    {
+      "maxLength": 64,
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Local source interval end using h/m/s units; timestamps retain the original source origin"
+}

addedInput schema / properties / execution_budget

Added value: +{
+  "anyOf": [
+    {
+      "additionalProperties": true,
+      "type": "object"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Explicit max_calls/max_tokens/max_output_tokens/max_frames/max_windows and start_ms/end_ms/fps. Bounded execution skips optional enrichment and cache reuse."
+}

addedInput schema / properties / fps

Added value: +{
+  "anyOf": [
+    {
+      "exclusiveMinimum": 0,
+      "maximum": 30,
+      "type": "number"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Local files only: requested static sampling frames per second; observed coverage remains unknown"
+}

addedInput schema / properties / output_fields

Added value: +{
+  "anyOf": [
+    {
+      "items": {
+        "type": "string"
+      },
+      "type": "array"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Select top-level response fields; complete source/citation/provenance carriers remain included"
+}

addedInput schema / properties / start_offset

Added value: +{
+  "anyOf": [
+    {
+      "maxLength": 64,
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Local source interval start, using h/m/s units, for example '27m' or '1m30.5s'"
+}

changedInput schema / properties / strict_contract / description

Previous value: -"Enable strict contract pipeline with quality gates, artifact rendering, and semantic validation. Produces richer output with strategy report, concept map, and HTML/Markdown artifacts."New value: +"Enable strict contract pipeline with quality gates, artifact rendering, and structural validation. Factual/media review and observed coverage remain explicitly pending or unknown. Produces richer output with strategy report, concept map, and HTML/Markdown artifacts."

addedInput schema / properties / transcript_limit

Added value: +{
+  "anyOf": [
+    {
+      "maximum": 10000,
+      "minimum": 1,
+      "type": "integer"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Page a top-level transcript string without truncating stored evidence or citations"
+}

addedInput schema / properties / transcript_offset

Added value: +{
+  "default": 0,
+  "description": "Original custom-schema transcript string offset in Unicode code points",
+  "maximum": 2147483647,
+  "minimum": 0,
+  "type": "integer"
+}

Changed3 schema fields changedv0.6.1May 26, 2026
- addedInput schema / properties / file_path / description
```
Added value: +"Path to a local video file."
```
- addedInput schema / properties / thinking_level / description
```
Added value: +"Gemini thinking depth."
```
- addedInput schema / properties / url / description
```
Added value: +"YouTube video URL."
```
First observedv0.6.0Mar 31, 2026

TDQS

A3.8/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Adds context beyond the annotations: Gemini structured output with default VideoResult schema, artifact rendering and quality gates under strict_contract, and the fact that factual/media coverage stays explicitly pending/unknown. It also discloses the latency tradeoff ('returns richer output but takes longer'), which explains why readOnlyHint=false despite it being an analysis call.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three short paragraphs, purpose and source constraint front-loaded, with the expensive strict_contract mode deferred to last. Every sentence carries information, though 'with any instruction' is mildly loose phrasing for a routing document.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 15-parameter tool with an output schema present, the description covers the decisions that matter (source exclusivity, output shaping, strict vs default pipeline). Remaining gaps (fps, offsets, execution_budget, transcript paging) are documented in the schema, so the definition is adequately complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, and the description earns above that by supplying a cross-parameter constraint ('exactly one of url or file_path') and the output_schema-vs-default-VideoResult rule that the schema alone does not state. It stops short of explaining the interval/budget/transcript paging params, which remain schema-only.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource ('Analyze a video') and covers both accepted source types (YouTube URL or local file) plus the open-ended 'any instruction' contract. It is clearly distinct from frame/storyboard/window siblings, though it never names a sibling explicitly to route the agent.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives a concrete usage constraint ('Provide exactly one of url or file_path') and a conditional mode ('When strict_contract=True...'), which is real guidance. However, it never says when to prefer this over video_analyze_windows, video_batch_analyze, or video_storyboard, so the alternative-selection guidance is only implied.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server