levea-mcp-server

Edit video (live progress)

autonomous_edit_streaming

Destructive

Executes long-running video edits and streams per-step SSE events as MCP progress notifications for live updates, then returns the final terminal result.

Instructions

Same as autonomous_edit but consumes the SSE event stream from the backend. Returns the final terminal result; per-step events (status, thinking, tool_call, tool_result, background_job_completed) are emitted as MCP progress notifications when the client supports them. Use for long-running briefs where users expect live progress.

Input Schema

Name	Required	Description
`scene`	No	Optional existing scene/timeline state to continue editing (advanced).
`assets`	No	Optional input media assets. URLs must be public HTTP/HTTPS links.
`prompt`	Yes	Natural-language description of the desired edit or creative brief. The backend planner decomposes it into canonical actions (viral clips, captions, reframe, export, …). Send the user's intent verbatim — do not pre-decompose.
`brandId`	No	Optional per-call brand profile id (see list_brand_kits).
`core_only`	No	Return rendering-essential scene fields only (smaller payloads).
`video_url`	No	Optional public source video URL (HTTP/HTTPS: S3, Google Drive public link, Dropbox, direct MP4) for a new project. Local file paths are not supported.
`editedPlan`	No	Optional pre-edited plan object returned by a prior awaiting_approval response.
`project_id`	No	Optional project id for tracking and continuity. Reuse the projectId returned by the first call for every continuation.
`flaggedIssues`	No	Optional list of specific user-flagged problems for this run (e.g. ["caption #3 misspelled"]).
`workingMemory`	No	Working memory returned by an awaiting-approval call; resend it to resume.
`attachedImages`	No	Optional reference image URLs used for planning.
`projectBrandId`	No	Optional default brand for the project. Persists; per-call brandId overrides.
`captionTemplateMode`	No	How to apply captionTemplatePreset.
`requirePlanApproval`	No	If true, the backend stops after planning and returns workingMemory for user approval. Resend with workingMemory (and optionally editedPlan) to continue.
`captionTemplatePreset`	No	Optional caption template to apply when (re)generating captions: a builtin slug (e.g. hormozi, mrbeast, viral-pop, karaoke, neon-pop, subtitle-bar) or a user-saved template id. Call list_caption_templates for the exact ids available.

Schema Changelog

Changes observed during successful MCP inspections.

Changed12 schema fields changedv0.6.2Oct 29, 2026

changedInput schema / properties / assets / description

Previous value: -"Optional asset descriptors. Each item: { id?, type: \"video\"|\"image\"|\"audio\", url|src, mimeType?, duration? }."New value: +"Optional input media assets. URLs must be public HTTP/HTTPS links."

addedInput schema / properties / assets / items / properties

Added value: +{
+  "duration": {
+    "type": "number"
+  },
+  "id": {
+    "type": "string"
+  },
+  "mimeType": {
+    "type": "string"
+  },
+  "type": {
+    "enum": [
+      "video",
+      "audio",
+      "image"
+    ],
+    "type": "string"
+  },
+  "url": {
+    "description": "Publicly accessible HTTP/HTTPS URL to the video, image, or audio asset.",
+    "type": "string"
+  }
+}

addedInput schema / properties / assets / items / required
```
Added value: +[
+  "url"
+]
```

changedInput schema / properties / attachedImages / description

Previous value: -"Optional base64-encoded reference images the agent should consider while planning."New value: +"Optional reference image URLs used for planning."

changedInput schema / properties / brandId / description

Previous value: -"Optional brand-profile id to apply for this call (palette, fonts, logo, voice, gradeBias). Look up ids via list_brand_kits. Overrides project default brand."New value: +"Optional per-call brand profile id (see list_brand_kits)."

changedInput schema / properties / captionTemplatePreset / description

Previous value: -"Optional caption template slug to apply when (re)generating captions. One of 41 builtins (hormozi, mrbeast, viral-pop, minimal-pro, karaoke, neon-pop, typewriter, pop-wave, subtitle-bar, creator-box, …) or a user-saved template id from list_caption_templates."New value: +"Optional caption template to apply when (re)generating captions: a builtin slug (e.g. hormozi, mrbeast, viral-pop, karaoke, neon-pop, subtitle-bar) or a user-saved template id. Call list_caption_templates for the exact ids available."

changedInput schema / properties / core_only / description

Previous value: -"If true, response carries only rendering-essential scene fields (no debug metadata)."New value: +"Return rendering-essential scene fields only (smaller payloads)."

changedInput schema / properties / project_id / description

Previous value: -"Optional project id for tracking / continuity."New value: +"Optional project id for tracking and continuity. Reuse the projectId returned by the first call for every continuation."

changedInput schema / properties / prompt / description

Previous value: -"Natural-language description of the desired edit or creative brief. The backend planner decomposes it into canonical actions (viral clips, captions, reframe, export, …). Send the user's intent verbatim — do not pre-decompose into multiple calls."New value: +"Natural-language description of the desired edit or creative brief. The backend planner decomposes it into canonical actions (viral clips, captions, reframe, export, …). Send the user's intent verbatim — do not pre-decompose."

changedInput schema / properties / requirePlanApproval / description

Previous value: -"If true, the agent stops after planning and returns { status: \"awaiting_approval\", workingMemory, plan } without mutating. Resume by re-calling with the returned workingMemory + an approval prompt (\"yes\", \"approve\", \"do it\")."New value: +"If true, the backend stops after planning and returns workingMemory for user approval. Resend with workingMemory (and optionally editedPlan) to continue."

changedInput schema / properties / video_url / description

Previous value: -"Optional source video URL. Auto-seeded as the starting layer when no scene is provided."New value: +"Optional public source video URL (HTTP/HTTPS: S3, Google Drive public link, Dropbox, direct MP4) for a new project. Local file paths are not supported."

changedInput schema / properties / workingMemory / description

Previous value: -"Working-memory snapshot returned by a prior call (typically one that returned awaiting_approval). Re-sending resumes that run."New value: +"Working memory returned by an awaiting-approval call; resend it to resume."

First observedOct 27, 2026

TDQS

A3.9/5.0

Behavior4/5

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

Annotations already flag this as a mutating, destructive, closed-world operation, so the safety burden is partly carried elsewhere. The description adds genuinely non-obvious behavior: it returns only the terminal result while emitting per-step events (status, thinking, tool_call, tool_result, background_job_completed) as MCP progress notifications, and only when the client supports them. It omits the approval/workingMemory continuation loop, which the schema covers.

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

Conciseness5/5

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

Three tight sentences, front-loaded with the differentiation from autonomous_edit, then return shape, then the selection rule. No filler.

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

Completeness3/5

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

For a 15-parameter, nested-object, mutating tool without an output schema, the description says almost nothing about what the "final terminal result" contains or how awaiting_approval / workingMemory / project_id continuity behaves. The rich schema compensates for parameters, but the return-shape gap keeps this at adequate.

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

Parameters3/5

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

Schema description coverage is 100%, so all 15 parameters, including nested scene/assets objects and the approval resume fields, are already documented in the schema. The description adds no parameter-level guidance at all, which is acceptable here but earns only the baseline.

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 and resource (edit video) and explicitly frames itself against the sibling autonomous_edit, with the distinguishing trait being SSE streaming plus progress notifications. It is slightly dependent on the reader already knowing what autonomous_edit does, but the delta is unambiguous.

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

Usage Guidelines4/5

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

"Use for long-running briefs where users expect live progress" gives a concrete selection condition and implicitly excludes short/synchronous edits (which map to autonomous_edit). It does not spell out failure modes or when streaming is unavailable, but the when-to-use signal is clear.

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

Edit video (live progress)

Instructions

Input Schema

Schema Changelog

TDQS

Other Tools