levea-mcp-server
This server lets an AI agent edit videos through Levea’s cloud editor, from natural-language end-to-end edits to precise timeline operations, project/asset management, and visual inspection.
Run autonomous prompt-based video edits synchronously, with live progress streaming, or as queued async tasks.
Make deterministic timeline edits: update, delete, or trim layers, and apply ~170 named transitions.
Discover and execute 189+ canonical editor actions and read/inspection tools individually or in atomic batches.
Render GPU-composited JPEG preview screenshots at chosen timestamps for visual QA before export.
List, apply, save, snapshot, and delete caption templates.
Create and manage brand kits with palettes, fonts, logos, voice tone, grade bias, and enforcement.
Manage projects and assets, including signed upload URLs, asset listing/deletion, and forced transcription.
Poll export jobs and queued edit tasks, and reconnect to a project’s active task.
Search and load production craft skills and read their bundled reference/template resources.
Check editor API connectivity with an unauthenticated health diagnostic.
Uses Google AI (Gemini) for multimodal planning, semantic analysis, and parameter parsing.
Integrates Google Cloud TTS for voiceover generation and Vertex AI for multimodal creative planning.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@levea-mcp-server Turn this creative brief and raw clips into an editable video project"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Levea — Prompt-First AI Video Editor (MCP Server & Cloud Production Studio)
Prompt-to-video editing without manual timelines. Describe the edit in plain language; Levea plans it, edits a real multi-layer project, checks the result on the rendered frames, and exports the MP4. An open-source MCP client connects your AI agent (Claude Desktop, Claude Code, Cursor, Cline, Gemini CLI, OpenClaw, Hermes, ChatGPT) to our cloud GPU video editor: frame-accurate cuts, 280+ caption templates, 2D and 3D motion graphics reviewed by an AI critic, a licensed music and sound-effects library, voiceovers, and active-speaker reframing, with nothing heavy running on your machine.
🚀 30-Second Quick Connect (Claude · ChatGPT · Claude Code · Cursor · OpenClaw · Hermes)
Where You Work | How to Connect Levea (Copy & Paste) |
Claude.ai (Web / Desktop / Mobile) | Settings → Connectors → Add custom connector → Paste |
ChatGPT (Web / Desktop Connectors) | Settings → Connectors → Create → MCP Server URL: |
Claude Code (CLI / IDE) | Run |
Claude Desktop (1-Click | Download |
Cursor / Windsurf / Cline | Add remote MCP |
OpenClaw / ClawHub |
|
NousResearch Hermes |
|
Agent Skills CLI ( |
|
Why Use Levea Inside Claude Code / Cursor Instead of Local ffmpeg, Python, or Standalone HTML/DOM Video Tools?
In Claude Code, an agent can run local ffmpeg scripts or render standalone HTML/CSS animations in headless Chrome. Here is the practical difference — and why Levea is a superset (Levea runs full 2D/3D motion graphics inside its cloud GPU NLE timeline alongside computer vision and GPU video compositing):
Capability | Raw | Standalone HTML/DOM Video Tools (Local Headless Chrome) | Levea MCP ( |
Best For | Simple container remuxing or fixed-coordinate trims | Standalone code-driven motion graphics, HTML decks & website promos | Real footage NLE editing + motion graphics + podcasts + viral shorts + VFX |
9:16 Active-Speaker Reframe | ❌ Crops dead-center ( | ❌ DOM/CSS has no per-frame face or mouth landmark detector | ✅ 68-point face & active-mouth tracking per frame + auto 2-up split-screen |
3D Camera Tracking, Hand Tracking & Rotoscoping | ❌ None | ❌ Browser DOM cannot solve 3D camera motion or track hands/masks on video | ✅ Monocular 3D camera solve, hand palm tracking, 2D region tracker, optical-flow warp stabilizer & AI alpha mattes |
Green Screen & GPU VFX Compositing | ❌ Basic | ❌ Headless Chrome captures DOM screenshots; no native video shader pipeline | ✅ Cloud GPU compositor: Keylight-grade despilled keying, depth bokeh, 328 GPU VFX presets, 170+ transitions, |
Kinetic Captions (280+ Templates) | ❌ Flat static | ⚠️ Must hand-code CSS per video; no automatic face/chin collision solver | ✅ 280+ GPU kinetic templates ( |
Multicam Podcast & Broadcast Audio | ❌ | ❌ Not an audio/multicam NLE; | ✅ Waveform multicam sync, ASR word-boundary cuts, filler removal, |
Motion Graphics | ❌ None | ✅ Full HTML/CSS/GSAP/Three.js compositions | ✅ Included natively — 2D/3D motion graphics plus After Effects shape layers (Trim Paths, Repeaters), composited over real tracked footage |
Human Final Polish ( | ❌ Must re-run 5-min | ❌ Must edit HTML/CSS code and re-render headless Chrome frames | ✅ 1-Click |
Where It Runs | Laptop only (CPU-heavy) | Laptop only (requires Node, Chromium, | Everywhere: Claude.ai Web/Mobile, ChatGPT, Claude Code, Cursor, Grok Bot, OpenClaw, Hermes (100% cloud GPU) |
Related MCP server: ViralSpin MCP
⚡ The Levea Philosophy: Prompt-to-Project
Unlike traditional one-shot AI video generators that output locked, un-editable pixels, Levea maintains a fully structured, multi-layer editable project and timeline containing video tracks, audio, text, masks, motion graphics, and brand kits.
Editable Multi-Track Projects: Every edit updates a persistent multi-layer timeline with full undo/redo revision history, so you can refine any cut, caption, or layer in follow-up prompts or open the project in the web studio.
Cloud GPU Rendering & Vision: Heavy computation (GPU rendering, transcription, active-speaker face tracking, 2D/3D motion graphics, and audio mastering) runs on our remote cloud GPU cluster. You get 4K and vertical reel exports without melting your local laptop CPU or installing CUDA.
Open-Source Client, Hosted Engine: The client-side MCP server wrapper (
levea-mcp-server), OpenClaw plugins, and skills are open-source (MIT). They connect to our hosted video production engine with OAuth 2.1 sign-in or an API key. Create a free account at livecore.ai.
🌐 Where to Use Levea (Every Client, Same Project)
Every Levea client connects to the same cloud editor and shares the same project workspace (edit contract). Start an edit from Claude, Cursor, ChatGPT, OpenClaw, Hermes, or Telegram, and open the exact same project in the web studio:
Client / Channel | How to Use |
Web studio | |
Telegram bot | @Levea_editing_bot — edit from your phone and link to your Levea account |
Levea HTTP API | API reference ( |
| npm · MCP Registry ( |
Hosted MCP (zero install) |
|
Claude Code plugin |
|
Claude.ai connector | Add |
ChatGPT app & Custom GPT | Connect |
Cursor plugin |
|
Gemini CLI extension |
|
Grok Bot | Add |
Mistral Le Chat | Add |
OpenClaw plugin | ClawHub ( |
Hermes plugin |
|
🛠️ Dual-Path Quickstart (Get Started in 60 Seconds)
Whether you are an AI developer looking to integrate automated editing into your agent loops, or a content creator building an automated faceless channel, Levea has a native path for you.
🟣 Claude Code plugin (one command)
Install Levea as a Claude Code plugin. It bundles the hosted Levea MCP server (OAuth sign-in, nothing to run locally) and a Levea editing skill:
/plugin install levea-video-editor --marketplace brajendrak00068/agentic-ai-video-productionOr from your shell:
claude plugin marketplace add brajendrak00068/agentic-ai-video-production
claude plugin install levea-video-editor@leveaOn first use, run /mcp, select levea, and sign in to your Levea account in the browser. Then describe the edit, for example: "Cut a 30-second vertical clip from https://example.com/talk.mp4 with word-by-word captions and export an MP4."
🌐 Hosted MCP (no install)
Any MCP client that supports remote servers can connect to the hosted endpoint directly:
https://api.livecore.ai/api/v1/misc/mcpIt speaks Streamable HTTP and signs you in with OAuth 2.1 (the client opens a browser to your Levea account). Clients that cannot do OAuth can send Authorization: Bearer <your Levea API key> instead.
Two hosts bundle this hosted server for one-step installs:
Gemini CLI extension — this repo is a Gemini extension (
gemini-extension.json):gemini extensions install https://github.com/brajendrak00068/agentic-ai-video-productionCursor plugin — add this repo as a plugin marketplace in Cursor (it reads
.cursor-plugin/marketplace.json), then installlevea-video-editor.
🤖 Use inside Grok Bot
Grok Bot can add an MCP server by URL, and the hosted server speaks the exact contract it expects (Streamable HTTP, OAuth 2.1 with RFC 7591 dynamic client registration). Tell your Bot:
"Add the Levea MCP server:
https://api.livecore.ai/api/v1/misc/mcp"
(or use Connect apps → add an MCP server by URL). A browser opens to sign in to your Levea account — then every Levea tool (autonomous_edit, queue_edit, direct canonical actions, job polling, caption templates, brand kits, projects, assets) is available to that Bot, and edits land in the same projects you see in the web studio. Clients that cannot do OAuth can use Authorization: Bearer <your Levea API key> instead.
🧑💻 Path A: For Developers & AI Engineers (The MCP Route)
Expose Levea as a client-side Model Context Protocol (MCP) server stdio wrapper (levea-mcp-server) in your favorite AI editors (Cursor, Cline, Windsurf, or Claude Desktop).
1. Get an API Key
Sign up at livecore.ai and generate a Levea API key.
2. Register the MCP Server
Add the following configuration block to your editor's MCP settings:
{
"mcpServers": {
"levea": {
"command": "npx",
"args": ["-y", "levea-mcp-server"],
"env": {
"LEVEA_API_URL": "https://api.livecore.ai",
"LEVEA_API_KEY": "your-key-from-livecore.ai"
}
}
}
}Note: LEVEA_API_URL should point to the bare host domain https://api.livecore.ai. The client-side wrapper handles endpoint appending automatically.
Client | Setup Commands / Instructions |
Cursor & Windsurf | Go to Settings -> Features -> MCP -> Add New MCP Server. Set type to |
Cline | Open settings, scroll to MCP, and add the config to the Cline MCP settings file. |
Claude Code | Run: |
Claude Desktop | Add the server block to |
Hermes | Native plugin: |
OpenClaw | Install via ClawHub: |
🎨 Path B: For Creators & Marketers (The OpenClaw Route)
Deploy Levea as an autonomous editing agent directly inside the OpenClaw workspace.
Install the Plugin: Search for
openclaw-ai-video-editoron ClawHub and install it. It adds the edit, queue, discovery, and status tools.Install the Skill: Add
levea-ai-video-editorfrom ClawHub Skills so your agent knows when and how to call the tools.Run your first edit:
"Review my vertical video, remove the opening silence, auto-apply Hormozi-style captions with highlighted keywords, add subtle background music, and export the Reels-ready MP4."
🧩 How Skills Work & How to Use Your Own Custom Skills
Levea is built around Skills (SKILL.md) at three levels — so your agent always edits with professional craft rules, and you can define your own reusable editing playbooks:
1. 100+ Built-In Production Craft Skills (On-Demand via MCP)
Levea ships with 100+ production craft playbooks covering short-form hooks, podcast repurposing, documentary montage, color grading & LUT finishing, kinetic caption typography, 2D/3D motion graphics, VFX keying & optics, broadcast audio mastering (-18 LUFS ducking), and lip-synced dubbing.
In Mode A (
autonomous_edit): Levea automatically selects and applies the matching craft skills for your prompt and footage.In Mode B (Direct Agent Control): Your agent can search, load, and follow any built-in craft skill directly in its own context window before calling
execute_actionorexecute_batch:search_skills— search all 100+ craft playbooks by keyword or category (e.g.query: "captions","color","podcast","montage").load_skill— load the fullSKILL.mdplaybook and its bundled resource list for one or moreskill_ids.read_skill_resource— read specific reference tables or templates (references/...,templates/...) belonging to a skill.
2. 12 Pre-Built Client Skills for Your Agent (skills/*)
This repository includes 12 ready-to-install agent skills under skills/ that teach local agents when and how to invoke Levea:
Skill | Focus |
Full end-to-end editing (Mode A & Mode B) | |
Multi-clip viral highlight extraction from long recordings | |
16:9 → 9:16 active-speaker vertical reframing & hooks | |
280+ word-synced caption templates, translations & dubbing | |
Waveform multicam sync & active-speaker switching | |
Word-boundary silence/filler cuts, voiceover & | |
Contextual B-roll cutaways, charts, lower thirds & callouts | |
Shape layers (Trim Paths, Repeaters), Bezier paths & 3D titles | |
Keylight green-screen keying, AI mattes, depth bokeh, LUTs & 328 VFX | |
2D region tracking, hand palm attachment, 3D camera solve & warp stabilizer |
Install all 12 skills in one command:
# Claude Code, Cursor, Codex, OpenCode, Gemini CLI
npx skills add brajendrak00068/agentic-ai-video-production
# OpenClaw / ClawHub
clawhub install levea-ai-video-editor3. How to Create & Use Your Own Custom Skills (SKILL.md)
You can write your own custom SKILL.md to lock in your channel's house style, brand kit (brandId), preferred caption template (captionTemplatePreset), reference video style (style_reference_url), and exact step-by-step editing recipe.
Where to save your custom skill:
Claude Code:
.claude/skills/my-show-style/SKILL.md(project) or~/.claude/skills/my-show-style/SKILL.md(personal)Cursor:
.cursor/skills/my-show-style/SKILL.mdor.cursor/rules/my-show-style.mdcOpenClaw:
~/.openclaw/skills/my-show-style/SKILL.md(or<workspace>/skills/my-show-style/SKILL.md)Hermes:
~/.hermes/skills/my-show-style/SKILL.mdOpenCode:
~/.config/opencode/skills/my-show-style/SKILL.md
Example Custom SKILL.md (Copy & Customize):
---
name: my-show-style
description: "Edit videos in our channel's signature 9:16 house style using Levea MCP."
---
# My Show House-Style Video Edit
When the user asks to edit a clip in our house style:
1. **Load built-in craft guidance (optional):**
Call `load_skill` with `skill_id: "caption-typography"` to combine Levea's typography rules with our house rules below.
2. **Option A — One-Call Autonomous Edit (`autonomous_edit`):**
Call `autonomous_edit` with:
- `brandId`: `"brand_your_kit_id"`
- `captionTemplatePreset`: `"hormozi"`
- `style_reference_url`: `"https://your-cdn.com/reference-reel.mp4"`
- `prompt`: "Cut silences >0.4s and filler words, reframe to 9:16 keeping the active speaker centered, add Hormozi captions in the lower safe zone, add a lower-third title card at 0:02, duck upbeat background music to -18 LUFS under speech, and export 1080p MP4."
3. **Option B — Deterministic Step-by-Step Recipe (`execute_batch` → `render_preview` → `EXPORT_VIDEO`):**
- Call `execute_batch` on `project_id` with our exact sequence:
1. `REMOVE_SILENCES` (`min_silence_duration: 0.4`, `padding_ms: 80`)
2. `CONFIGURE_CANVAS` (`aspect_ratio: "9:16"`)
3. `VERTICAL_REFRAME` (`mode: "active_speaker"`)
4. `COLOR_GRADE` (`preset: "cinematic_warm"`, `contrast: 1.08`, `saturation: 1.05`)
5. `GENERATE_CAPTIONS` (`template: "hormozi"`)
6. `AUTO_DUCK` (`target_lufs: -18`)
- Call `render_preview` at `timestamps: [1.5, 5.0, 10.0]` to visually inspect the GPU-rendered JPEG frames and verify captions do not overlap faces.
- Call `execute_action` with `action: "EXPORT_VIDEO"` to deliver the final MP4.Once saved, simply tell your agent: "Edit https://.../episode12.mp4 using my-show-style" — your agent reads your custom SKILL.md and drives Levea with your exact rules every time.
🧩 OpenClaw Integration: Plugin vs. Skill
Levea is built for modular agent platforms. When deploying Levea inside platforms like OpenClaw or Hermes, we split execution capabilities from cognitive strategies:
The Plugin (
openclaw-ai-video-editor): the tools. It registers the edit, direct action, skill discovery, queue, and status tools with their schemas, manages the API key, and streams progress.The Skill (
levea-ai-video-editor— or your own customSKILL.md): the instructions. It tells the agent when to use Levea, which parameters or house-style rules to pass, and how to handle approvals, visual preview checks, long-running edits, and follow-ups.
We recommend installing both to unlock the full power of autonomous editing with a professional finish.
🎬 What It Does: Say It in Plain Language
Say it in plain language; Levea plans, executes, and finishes the complete edit:
You ask | It does |
"Turn this into 5 viral clips with captions and vertical reframe" | Identifies narrative peaks, cuts clips, applies bold captions, reframes to 9:16 vertical, and exports. |
"Cut a 60-second highlight from this 2-hour podcast" | Discovers strongest continuous story segment, trims dead air, and packages with captions. |
"Make this TikTok-ready" | Vertical 9:16 reframe + captions + silence removal + keyword emphasis kit. |
"Export for TikTok, Reels, Shorts, YouTube, and Instagram" | One pass, all platform-specific aspect ratios in one package. |
"Replace the green screen with a beach, keep the speaker centered" | Chroma key + background composite + face tracking in one call. |
"Remove the background — no green screen" | AI background removal via an alpha matte on any footage. |
"Swap my background for a city skyline — no green screen" | AI matte isolates subject cleanly, compositing the new background behind. |
"Remove all silences and filler words, add background music" | Cleans audio track (cuts |
"Auto-zoom on whoever's talking" | Active-speaker detection + dynamic zoom-follow framing. |
"Caption this and highlight every time they say 'launch'" | Auto-captions + keyword emphasis (scaling / glow / pulse). |
"Find every clip where Alex appears" | Face identity search across the clips in your project. |
"Add narration in a cloned voice over the intro" | Voice cloning + TTS overlay + auto-ducking. |
"Caption this Hormozi-style with karaoke word highlighting" | Word-synced karaoke captions — active word fills and underlines in sync with speech. |
"Generate B-roll over the product mention" | AI B-roll generation + placement at the exact timestamp. |
"Color grade this like a Netflix doc" | Cinematic LUT color grade with contrast and film-look curve. |
"Slow-mo the climax, freeze on the reveal" | Speed change ( |
"Reframe to vertical but don't crop the lower-third captions" | Caption-safe 9:16 reframe (detects on-screen text regions and reframes around them). |
"Pull the key stats from this and animate them as charts" | Transcript-driven SVG charts (bars, lines, donuts) + animated stat-callout motion graphics. |
"Sync these 3 camera angles and cut between them on the active speaker" | Multi-cam audio cross-correlation sync + automatic angle switching. |
"Blur the license plates and bleep the swearing" | Privacy redaction (face and moving object blur) + transcript profanity cleanup. |
"Turn this YouTube video into a 30-second reel with motion graphics that never cover his face" | Hook-first trim, face-tracked 9:16 reframe, word captions, motion graphics placed clear of every face, ducked music. |
"Make a 15-second 3D animation of a little robot racing a deadline" | An original 3D character film with camera moves, lighting, music, and sound design, reviewed by an AI critic before delivery. |
"Add upbeat music from the library and a whoosh on every cut" | Licensed music fitted to the video length, sound effects placed on the cuts, voice-aware ducking. |
"Speed-ramp into the jump and use J-cuts between speakers" | Speed ramps / time remap and J/L, match, and smash cuts. |
"Push the camera in slowly with a little parallax" | Keyframed 3D camera moves with depth-based parallax. |
"Match the color of the second camera to the first" | Auto color and shot-to-shot color matching, plus film looks, film burn, and lens flares. |
"Dub this into Spanish in my voice and lip-sync it" | Translated dub in a cloned voice, lip-synced to the speaker's mouth. Mouth-to-audio sync is measured and verified, and the dub is checked by listening to it (right language, intelligible). Lines that had to be shortened to fit are reported. |
"Put the price tag in her hand and keep it there" | Hand tracking: the overlay follows the palm, can grow and shrink with the hand, and hides while the hand is out of view. |
"Mix these four songs into one continuous DJ set" | Beat-matched, harmonically mixed program: each track's tempo and musical key are measured, tracks are time-stretched and key-shifted (Camelot wheel) so they mix. |
"Cut a reel from this explainer and keep the slides readable" | Slides, diagrams, and designed graphics are cropped to their content and shown with a headline on a matched background instead of a tiny letterboxed strip. |
⚡ Levea vs. Legacy Editors
Capability | Legacy (Premiere, DaVinci, CapCut, Descript) | Levea Agentic Video Editor |
Interface | Drag, drop, keyframe by hand on a manual timeline | One natural-language prompt or direct tool call; the agent finishes the edit |
Auto-analysis on upload | Manual scene detection + subtitles | Faces, active speakers, shot cuts, and on-screen text regions detected automatically |
"Make this viral" | You manually hunt for hooks, crop, splice, and align | Single preset — vertical 9:16 + captions + silences + face tracking |
Face search | Filename search | "Find every clip where Alex appears" across the clips in your project |
Background replace | Key out green screen or draw rotoscope masks by hand | One call — an AI alpha matte handles any footage without a green screen |
Broadcast-grade audio | Manual loudness metering + compression + EQ | Auto LUFS loudness-normalized (EBU R128), true-peak limited, music ducked under speech |
Editorial reasoning | You listen and hunt for the climax | Agent surfaces narrative peaks and high-retention moments |
Verification & Repair | You eyeball it and manually re-render | Every edit is checked on the rendered frames; failures are repaired automatically and reported with a typed verification report |
Multi-platform export | One render per aspect ratio | TikTok + Reels + Shorts + YouTube in one pass |
Extensibility | Plugins call external binaries | 37 MCP tools unlocking 189+ editing actions, 13 inspection tools, and 100+ craft skills |
🎬 Creator & Dev Use Cases
Use case | Typical production path |
Faceless channel | Segment the transcript, plan the layout, overlay B-roll, add library or generated music, and render. |
Auto-captions | Transcribe, time every word, highlight keywords, and style with a caption template. |
Shorts & Reels highlights | Find the hook, trim to the target length, reframe to 9:16 around the speaker, add captions and motion graphics. |
Interview polish | Cut long silences, bleep profanity, grade color, and add lower-third speaker graphics. |
Green screen & backdrop swap | Key or matte the speaker, layer a new background photo or video, and composite. |
Multicam active-speaker cuts | Sync camera angles by audio, identify speakers, and cut to whoever is talking. |
⚙️ Verification & Repair
Levea checks its own work before it reports success.
Every step is verified when it runs: checks such as "the layer exists", "the timing matches the request", "no graphic covers a face", or "captions stay in sync". A failed check names what was expected and what was found.
Automatic repair fixes what failed: it applies targeted fixes and verifies again until the check passes or the request budget runs out, without destroying the user's own layers or breaking earlier steps.
A final review looks at the rendered frames of the finished edit and can make further fixes.
The agent inspects the result the way an editor would: it watches the rendered video, listens to the rendered audio (is the dub in the right language, intelligible, on-content?), reads color scopes (crushed blacks, clipped whites, flat saturation), previews the frame inside the TikTok / Reels / Shorts interface to check nothing sits under the app's buttons, and measures lip sync on dubbed footage.
The result says what happened:
verificationReport.statusisverifiedonly when that final review approved the edit;unverifiedmeans no verdict was reached (the edit is kept, never reported as verified), with every failed check and what repair tried.Corrections are a follow-up prompt with the same
project_id. Every revision is saved, so a change can be undone.
📂 Capability Status
Availability of specific tracks varies by deployment, active model tiers, and account quotas.
Supported Production Paths
Project and timeline state: Full NLE track and clip operations (insert, overwrite, lift, extract, ripple delete/trim, slip, slide, freeze frame, reverse), multi-track management, nesting containers, and durable linear undo/redo.
Captions: automatic captions, word timing, keyword emphasis, 280+ caption templates (Hormozi, Spotlight, Karaoke, Typewriter, phrase lockups, two-line podcast looks, …), animated word entrances and burn-aways, and Latin, Indic, Arabic, Cyrillic, and Greek text.
Motion graphics: lower thirds, title cards (59 templates), real 3D extruded bevelled titles, charts, counters, flow diagrams, collage layouts, any-layer behind-the-subject occlusion, native After Effects shape layers (arc-length Trim Paths, Repeaters, shape modifiers — pucker & bloat, zig zag, wiggle paths, round corners, offset paths), spatial Bezier motion paths with auto-orient, graph-editor velocity/influence easing, echo motion trails, and full motion-graphic films (2D motion design and 3D character films with camera, lighting, music, and sound design). Every graphic is placed clear of faces and reviewed by an AI critic before it ships.
Editing techniques: speed ramps, time remap, and posterize time, J/L, match, twitch, and smash cuts, keyframed 3D camera moves with depth parallax, non-destructive warp stabilization (smooth camera path + border-hiding auto-scale), AI outpainting to new aspect ratios, eye-contact correction, face retouch, style packages, and 170+ transitions.
Layout and perception: multi-cam synchronization and active-speaker cuts, scene and shot cut analysis, face detection and tracking, face identity search across a project's clips, on-screen text-region detection and OCR recognition, safe zones, 2D feature-region tracking (position, rotation, and scale), live null parenting, planar corner-pinning, monocular 3D camera solve, and hand tracking (attach a sticker, label, or prop to a hand's palm).
Compositing and color: Keylight-grade screen color key (with despill, clip black/white, softness, shrink/grow), Lab/YUV/RGB color range key, dedicated spill suppression, SVG Bezier masks, feathered/inverted mattes, track mattes, blend modes (all 29 After Effects modes), adjustment layers, ordered multi-effect stacks (up to 4 chained effects per layer), alpha-matte background replacement, on-screen text / logo / watermark removal, auto color and shot matching, monotone cubic curves, CMYK per-range selective color, 5-point levels, LUTs (including your own
.cube1D/3D files), After Effects two-tone threshold glow, depth-map-driven camera lens blur (polygonal iris bokeh), optical lens flares, and 328 procedural VFX presets (film burn, bloom, rain, smoke, fire, glitch, portal, lightning, …).Audio: a reviewed, licensed music and sound-effects library, sound design (effects on cuts and beats, room tone), music fitted to the video length, silence and filler-word cleanup, word-level muting, crossfades, EQ, denoise, EBU R128 loudness normalization, voiceover, voice cloning and voice conversion, speech-aware ducking, beat- and drop-synced edits, measured music structure (sections and drops), musical key detection, DJ mixes (beat-matched, harmonic track-to-track mixing), vocal / instrumental separation, and lip-synced dubbing (translated voice in the speaker's cloned voice, with measured mouth-to-audio sync).
Autonomous Directing: automated cold-open hook extraction, hook-first trims to a target length, podcast polish macros, contextual B-roll placement, auto lower thirds, slideshow generation, and reels cut from explainer videos with slides and diagrams cropped to their content.
Sources and delivery: public video links (including Google Drive files and folders) checked before planning (Drive files that hit Google's download quota are fetched in the background instead of failing), and MP4 export in every platform aspect ratio.
Verification: checks on the rendered frames, requirement tracking, the agent watching and listening to the rendered result (color scopes, platform-UI preview, audio-content and lip-sync checks), brand kits with hard enforcement (off-palette colors fail the check and are repaired), an automatic repair loop, plan approval (
requirePlanApproval+planDecision), and a typedverificationReporton every result.
Model- or Deployment-Dependent
Generated video, images, B-roll, music, sound effects, voiceover, and voice cloning quality/availability vary by model tier and quota (all are wired as canonical actions in production).
Neural alpha matting and background replacement quality.
Lip-sync dubbing quality (a lip-sync step passes only on its measured mouth-to-audio sync).
CJK caption rendering: right-to-left shaping ships in production; CJK glyph coverage depends on font fallback.
📂 MCP Tool Surface (37 Annotated Tools Unlocking All 189+ Canonical Actions, 13 Read Tools & 100+ Craft Skills)
The MCP server (levea-mcp-server, 37 tools) exposes both high-level goal delegation (autonomous_edit) and direct progressive-discovery & execution (search_tools, get_action_schema, execute_action, execute_batch, render_preview, search_skills, load_skill, read_skill_resource, plus update_layer, delete_layer, trim_layer, apply_transition), alongside typed management and polling tools. The OpenClaw and Hermes plugins expose the same dual-mode editing and discovery surface.
Group | Tools | Description |
Mode A: Goal Delegation |
| Send a natural-language brief; Levea's autonomous editing agent inspects the footage, edits, verifies on rendered frames, and exports. |
Mode B: Universal Direct Canonical Suite (189+ Actions, 13 Read Tools & 100+ Skills) |
| Discover any of the 189+ canonical actions, 13 read tools, and 100+ editing craft skills on demand, execute single actions or atomic batches (with automatic safe-zone checks, timeline reconciliation, and async asset completion), and inspect real GPU-composited JPEG screenshots ( |
Deterministic Layer Shortcuts |
| Fast typed shortcuts for single-layer property updates, deletes, trims, and ~170 named transitions. |
Job Polling |
| Track rendering, generation jobs, and the active task of a project. |
Caption Templates |
| Manage styling templates (280+ built-in: Hormozi, Spotlight, Karaoke, …); |
Brand Kits |
| Manage brand colors, fonts, logos, voice tone, and grading rules. |
Projects |
| Project workspace management. |
Assets |
| Signed upload URLs, asset listing and deletion, and transcription. |
Diagnostics |
| Unauthenticated connectivity check. |
When to Use Mode A (autonomous_edit) vs. Mode B (execute_action / execute_batch / render_preview)
Scenario | Best Mode | Why |
Broad creative brief / full video from prompt | Mode A ( | Autonomous editing agent + 100+ craft skills handle decomposition, pacing, and verification in 1 call |
Exact parameter tweak, custom timestamp math, or local-file-driven script | Mode B ( | Instant (<50ms), zero backend LLM cost, 100% deterministic control over all 189+ actions |
Visual QA loop (like Claude Code | Mode B ( | Returns real GPU-composited JPEG screenshots ( |
Heavy first draft + surgical fine-tuning | Hybrid (Mode A → Mode B) | Fast first draft via |
Every edit result carries a verificationReport (verified, rejected, unverified, or not_applicable, plus each requirement that failed its check and what repair tried) and qualityNotes.
🔮 FAQ
How does Levea compare to generic video generators like Sora, Veo, or Runway?
Generic generative video models (Sora, Runway, Veo) output raw, locked pixels. You cannot edit a layer, adjust caption typography, swap background music, or correct a word timing afterward. Levea is a full-featured timeline editor that builds structured project layers. It uses generative models only as optional asset generation steps, keeping your editing pipeline fully editable, inspectable, and deterministic.
Is Levea safe to use for client and enterprise work?
It is designed for it. Every mutation runs through a typed operation engine, and every action the planner or the repair loop chooses is authorized against the request before it runs, so an LLM cannot inject arbitrary mutations. If you set requirePlanApproval: true, the system halts and returns the creative plan for approval; resend with planDecision: "approve" (or an editedPlan) to run it.
What coding and agent environments does Levea support?
Levea integrates natively with Model Context Protocol (MCP) hosts like Claude Desktop, Cursor, Cline, Windsurf, Claude Code, and the Gemini CLI (via the bundled extension). For standalone agent systems, Levea exposes structured packages for OpenClaw and Hermes.
Can I use custom brand fonts, logos, and specific caption templates?
Absolutely. Using our Brand Kits API and tools, you can declare custom typography scales, palette hex codes, logo image references, and voice tone. The planning model reads these rules and automatically enforces them across the timeline during the composition pass.
🔗 Links & Resources
Levea Studio and API Keys: livecore.ai
npm MCP Server Wrapper:
levea-mcp-serverMCP Registry:
io.github.brajendrak00068/levea-mcp-serverOpenClaw Plugin Page:
openclaw-ai-video-editorOpenClaw Skill Page:
levea-ai-video-editorDetailed Agent Integration Guide: AGENTS.md
MCP Folder Documentation: mcp-server/README.md
How long an edit takes: read-only questions answer in seconds, a simple edit with export in a few minutes, and a full edit with motion graphics, music, and export in 10–20 minutes. Clients return status: "running" with the project_id if they stop waiting first; the edit keeps running (see AGENTS.md).
Privacy & telemetry: edits send your prompt, source links, and project ids to the Levea API (privacy policy, terms). The MCP server and OpenClaw plugin also send a tiny anonymous usage ping (event name, client version, whether a key is set; never prompts, media, or keys); turn it off with LEVEA_TELEMETRY=off.
Security: see SECURITY.md. Contributing: see CONTRIBUTING.md.
Support & Contact: brajendra@livecore.ai · GitHub issues
License
Available Tools
37 toolsapply_caption_templateApply caption templateAInspect
Apply a caption template to a scene synchronously (without going through autonomous_edit).
Inputs: template_id (required), mode, scene, project_id.
apply_transitionApply transitionAInspect
Deterministic edit: apply a named transition between two adjacent clips, committed immediately (the same operation as the editor's transition picker). type is one of ~170 names, e.g. fade, cross_dissolve, dip_to_black, slide_left, push_left, wipe_left, zoom_in, whip_pan, light_leak, glitch, zoom_blur, page_curl, crt_power_off. An unknown name is rejected with the reason. Both layer ids must exist in the project. object_match_cut and parallax_occlusion need face tracking and are refused here — ask autonomous_edit for those.
Inputs: type (required), project_id (required), to_layer_id (required), easing, +3 more.
asset_upload_urlGet asset upload URLAInspect
Get a signed upload URL for a local file (so users do not have to host video / image / audio elsewhere before passing video_url to autonomous_edit). Body: { fileName, fileType?, contentType?, projectId? }. Returns { uploadUrl, fileUrl }; PUT the file bytes to uploadUrl, then pass fileUrl to autonomous_edit as video_url / in assets[].
Inputs: fileName (required), fileType, projectId, contentType.
autonomous_editEdit videoADestructiveInspect
PRIMARY EDIT TOOL. Run a natural-language video edit end-to-end. The backend agent plans, executes, verifies, and exports. Accepts anything from a one-liner ("generate 3 viral clips with captions and export") to a multi-paragraph creative brief. Returns the final mp4 URL synchronously when an export is queued. For long renders you may prefer queue_edit + check_job_status.
Inputs: prompt (required), scene, assets, brandId, +13 more.
autonomous_edit_streamingEdit video (live progress)ADestructiveInspect
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.
Inputs: prompt (required), scene, assets, brandId, +13 more.
check_job_statusCheck job statusARead-onlyInspect
Poll an export / async job by jobId (numeric). Supports wait:true to block until terminal.
Inputs: job_id (required), wait.
check_task_statusCheck task statusARead-onlyInspect
Poll a queued edit task by taskId (returned by queue_edit). Separate from job_id (exports).
Inputs: task_id (required).
create_brand_kitCreate brand kitBInspect
Create a brand profile. Fields: name, palette {primary,secondary,accent,neutral,extras[]}, fonts {primary,secondary}, gradeBias (warm/cool/neutral/vivid/muted), forbiddenColors[], logoAssetUrl, voiceTone, enforcement (soft|hard).
Inputs: name (required), fonts, palette, gradeBias, +4 more.
create_projectCreate projectCInspect
Create a new project. Optional initial brandId.
Inputs: name (required), brandId, description.
delete_assetDelete assetADestructiveInspect
Delete an asset by file name (the path returned by asset_upload_url).
Inputs: fileName (required).
delete_brand_kitDelete brand kitCDestructiveInspect
Delete a brand profile.
Inputs: brand_id (required).
delete_caption_templateDelete caption templateBDestructiveInspect
Delete a user-saved caption template by id.
Inputs: template_id (required).
delete_layerDelete layerADestructiveInspect
Deterministic edit: permanently delete one or more named layers of a project, committed immediately. close_gap: true ripples the timeline closed and works with a single layer_id only. Selectors such as "all" or "@…" are refused, and so is a delete that would leave the project empty. To hide a layer temporarily use update_layer with {"visible": false} instead.
Inputs: project_id (required), layer_id, close_gap, layer_ids.
delete_projectDelete projectDDestructiveInspect
Delete a project.
Inputs: project_id (required).
editor_healthCheck editor healthARead-onlyInspect
Connectivity diagnostic against LEVEA_API_URL. Unauthenticated.
Inputs: no named parameters.
execute_actionExecute canonical editor actionADestructiveInspect
Execute any single canonical editor action (all 189+ actions) or read/inspection tool (list_layers, inspect_layer, query_transcript, list_sources, search_shots, shot_timeline, read_source_transcript, measure_audio, measure_picture, measure_color) directly on a project scene without autonomous_edit. Pass video_url or assets on the first call to bootstrap a new project.
Inputs: action (required), scene, assets, video_url, +3 more.
execute_batchExecute canonical action batchADestructiveInspect
Execute an ordered array of canonical actions (and/or read tools) in a single atomic round-trip without autonomous_edit. Automatically runs safe-zone checks, timeline reconciliation, layout de-stacking, async asset completion, and commits the updated scene to project_id.
Inputs: actions (required), scene, assets, video_url, +2 more.
get_action_schemaGet canonical action schemaARead-onlyInspect
Fetch the full JSON inputSchema, parameter constraints, and description for any canonical editor action (e.g. COLOR_GRADE, TRIM_LAYER, GENERATE_CAPTIONS, PLAN_MOTION_GRAPHICS, RENDER_PREVIEW, EXPORT_VIDEO) or read tool (e.g. list_layers, query_transcript, inspect_layer).
Inputs: action (required).
get_active_taskGet active taskARead-onlyInspect
For a given project, return the currently-running edit task (if any). Useful for reconnecting MCP clients.
Inputs: project_id (required).
get_brand_kitGet brand kitCRead-onlyInspect
Fetch one brand profile by id.
Inputs: brand_id (required).
get_projectGet projectCRead-onlyInspect
Fetch one project by id.
Inputs: project_id (required).
list_assetsList assetsBRead-onlyInspect
List the user's uploaded assets. Optional projectId filter.
Inputs: fileType, projectId.
list_brand_kitsList brand kitsARead-onlyInspect
List the user's brand profiles (palette, fonts, logo, voice, gradeBias). Returns ids you pass to autonomous_edit's brandId / projectBrandId.
Inputs: no named parameters.
list_caption_templatesList caption templatesARead-onlyInspect
List the user's caption templates (builtin + saved). Returns slugs, display names, and style summaries.
Inputs: no named parameters.
list_projectsList projectsCRead-onlyInspect
List the user's projects.
Inputs: no named parameters.
load_skillLoad production craft skillARead-onlyInspect
Load the full craft playbook (SKILL.md) and bundled resource list for one or more skill IDs returned by search_skills.
Inputs: skill_id, skill_ids.
queue_editQueue video editADestructiveInspect
Fire-and-forget async path. Returns immediately with a taskId. Use when the edit may exceed the MCP request timeout (long viral-clip generation, multi-format export). Poll with check_task_status or get_active_task.
Inputs: prompt (required), scene, assets, platforms, +5 more.
read_skill_resourceRead skill reference resourceARead-onlyInspect
Read a bundled reference or template file (e.g. "references/..." or "templates/...") belonging to a production skill.
Inputs: resource (required), skill_id.
render_previewRender GPU preview screenshotsARead-onlyInspect
Render real GPU-composited JPEG screenshots of the live project timeline at specified timestamps (and optional normalized 0..1 region crop) — the Levea equivalent of ffmpeg frame extraction, showing all video layers, captions, title cards, color grades, masks, and VFX shaders. Returns native MCP image blocks so you can visually inspect the frames before calling EXPORT_VIDEO.
Inputs: scene, region, project_id, timestamps.
save_caption_templateSave caption templateBInspect
Save a new caption template definition the user can re-apply later.
Inputs: name (required), style (required), tags.
save_current_caption_templateSave current captions as templateBInspect
Snapshot the caption styling currently applied to a project / scene as a reusable template.
Inputs: name (required), scene, project_id.
search_skillsSearch production craft skillsBRead-onlyInspect
Search Levea production craft skills (SKILL.md playbooks across editing, color grading, captions, audio mixing, motion graphics, VFX, documentary, podcast, and social shorts) by keyword or category.
Inputs: limit, query, category.
search_toolsSearch canonical editor toolsARead-onlyInspect
Search all 189+ canonical video editing actions, 13 read/inspection tools, AND canonical preset catalogs (280+ caption templates, 328+ GPU VFX effects, 170+ transitions, 50+ color grade LUTs, 59 title cards, fonts, blend modes) by keyword, category, or catalog filter. Returns matching actions, catalog presets, and the 5-phase NLE editorialPlaybook.
Inputs: limit, query, catalog, category.
transcribe_assetTranscribe assetBInspect
Force transcription of an existing asset. Returns transcript words + timing.
Inputs: fileName (required), language.
trim_layerTrim layerADestructiveInspect
Deterministic edit: shorten a layer, committed immediately. Either a relative trim (edge "start" | "end" + seconds to remove) or an absolute range in TIMELINE seconds (start_time + end_time: the part of the layer's current window to keep). On a single-source cut timeline the trim ripples later segments closed. A trim that removes nothing is an error (NO_CHANGE).
Inputs: layer_id (required), project_id (required), edge, seconds, +2 more.
update_brand_kitUpdate brand kitCDestructiveInspect
Update an existing brand profile by id. Body has the same shape as create_brand_kit.
Inputs: brand_id (required), name, fonts, palette, +5 more.
update_layerUpdate layerAInspect
Deterministic edit: set registry properties on one existing layer of a project, committed immediately (no planner, no export). E.g. {"x": 540, "y": 960}, {"scale": 1.2}, {"opacity": 0.5}, {"volume": 0.3}, {"visible": false}, {"z_index": 4}, or its timeline window {"start_time": 2, "end_time": 8}. Unknown keys are ignored, and an update that changes nothing is an error (NO_CHANGE). Not for keyframed animation (use autonomous_edit), and not "text" on a caption group (refused: it would overwrite every caption). Get layer ids from get_project (project.scene.layers[].id).
Inputs: layer_id (required), project_id (required), properties (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 37 tools
Descriptions do a lot of work to separate overlapping paths (autonomous_edit vs autonomous_edit_streaming vs queue_edit, execute_action vs execute_batch vs the deterministic update_layer/trim_layer/apply_transition tools), but there are still several routes to the same outcome, so an agent must reason carefully to pick one. get_action_schema, search_tools, and execute_action also partially overlap in discovery/schema-fetching roles, and apply_caption_template vs save_current_caption_template vs apply_caption_template can be confused at a glance.
Nearly all names follow a predictable snake_case verb_noun pattern (list_projects, create_brand_kit, delete_layer, check_job_status). A few outliers break the verb-first convention (editor_health, asset_upload_url, autonomous_edit_streaming), but they remain readable and unambiguous.
At 37 tools the surface is heavy for a single server, and several earn only marginal place (autonomous_edit_streaming largely duplicates autonomous_edit, get_action_schema overlaps search_tools, and caption-template save/snapshot are split into two tools). The broad domain (editing, brand kits, assets, skills, jobs) explains some breadth but does not justify the full count.
Coverage is strong: full CRUD for projects and brand kits, complete caption-template lifecycle, asset upload/list/delete/transcribe, job and task polling, and both high-level agentic and deterministic edit paths plus skills discovery. Minor gaps exist — no project update/settings operation and no get_caption_template detail fetch — but these are workaroundable.
Maintenance
Related MCP Connectors
AI video editor for agents and humans: timeline, captions, color, audio and generation as MCP tools.
The agentic video editor for ChatGPT and Claude. Ask in your chat and it cuts, captions, adds music, titles and motion graphics on a real timeline you can open and adjust.
AI video editor your agent can drive: plan scenes, edit a multi-track timeline, and render MP4s.
Compose, edit and render video with your AI agent. Turn footage into finished edits with content-aware reframing, word-timed captions and motion graphics. Hosted MCP with OAuth sign-in; cloud processing uses credits.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to edit videos through natural language, providing tools for timeline editing, audio management, rendering, and more.2MIT
- AlicenseAqualityDmaintenanceEnables AI agents to edit video assemblies from A-roll and B-roll, add captions, and publish to social media platforms.27MIT
- FlicenseNot gradedqualityBmaintenanceEnables AI agents to read, edit, and export real video projects with a full multitrack timeline, effects, captions, and export capabilities.2-
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to autonomously edit raw video footage into publish-ready videos with millisecond-accurate cuts, Whisper transcription, karaoke captions, motion graphics, and high-CTR thumbnails.7 npm6MIT