Use when generating or editing video with Runway models on Scenario via MCP: text-to-video or image-to-video from a first frame with Gen4.5 (cinematic motion, sequenced actions, camera choreography), or video-to-video editing of real footage with Aleph 2 (remove objects, swap products, restyle or relight shots, keyframe-pinned reference images, motion and audio preserved). Keywords: Runway ML, Gen4.5, Gen-4.5, Aleph, T2V, I2V, V2V, first frame.
中文简介正在准备中。当前显示原文说明。
实测评分 未实测
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Runway's family on Scenario splits by direction: Gen4.5 generates new clips from text or a
first-frame image; Aleph 2 rewrites existing footage, changing what the prompt names while motion,
audio, and structure carry over. Discover both with search; model_schema_get is the contract.
Connection and the core loop: the scenario skill; model-agnostic video work: the scenario-video
skill. If a sibling skill named here is missing from your available skills, ask the user to install
it (npx skills add scenario-labs/skills --skill <name>); unattended, proceed from tool schemas and
flag the gap.
Mode follows from the inputs (names and caps from the live schema):
| Mode | Inputs | Behavior |
|---|---|---|
| Text (Gen4.5) | prompt | aspectRatio honored; duration 2 to 10 seconds |
| First frame (Gen4.5) | image + prompt | opens on that frame; same duration; aspectRatio ignored |
| Edit (Aleph 2) | video + prompt | rewrites visuals; motion, audio, structure persist |
At authoring time Aleph capped sources at 30 seconds and 16 MB, separate limits: high-bitrate clips
hit the size cap first. Its median cost ran six times Gen4.5's, a fifteen-fold spread following the
source video: dry_run the exact payload first. Median latency neared six minutes, so jobs_wait
will time out mid-job; re-call it with pending_job_ids.
Text mode needs both the visuals and the motion described. A first frame fixes subject, composition,
and palette: spend the whole prompt on motion, camera, and atmosphere. Order beats explicitly
("kneels, then rises, camera cranes up") and give multi-beat prompts 8 to 10 seconds of duration.
Name camera moves plainly (locked camera, pan, dolly, tracking, orbit, crane), one or two per clip,
in present-tense prose.
aspectRatio takes pixel pairs ("1280:720", "960:960"), and text mode honored only the 16:9 and
9:16 pairs at authoring time; get other shapes (1:1, 4:3, 3:4, 21:9) from a first frame with that
geometry instead.
An Aleph prompt states the edit: remove, swap, restyle, relight. Everything unnamed persists, along
with motion and audio: Aleph cannot re-choreograph or add camera moves; when motion must change,
generate with Gen4.5. To steer the edit, pin up to 5 reference images: keyframes places each image
(uri) by seconds or at (a 0 to 1 fraction); promptImage by positionType plus
timestampSeconds or positionPercentage. The arrays are alternatives, never combined.
search with target="models", query="runway", public=true, e.g. model_runway-aleph-2 for
editing, model_runway-gen4-5 for generation (live hits at authoring time: re-discover each
session).model_schema_get with the Aleph id before anything else.scenario-video); upload_asset the clip and
product still (see the scenario skill).model_run with dry_run=true and the exact
parameters={"prompt": "Replace the soda can with the bottle from the reference image. Keep everything else unchanged.", "video": "asset_clip", "keyframes": [{"uri": "asset_bottle", "at": 0}]};
re-estimate after any trim.wait=false, then jobs_wait the returned job id, re-calling with
pending_job_ids; a timeout is not a failure and never justifies a second model_run.asset_display the output; check the swap holds in every shot.keyframes and promptImage: exclusive alternatives; pick one.aspectRatio as "16:9" or alongside image: pixel pairs only, and a first frame
overrides it.还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Handle review comments received on GitHub for the current branch's PR or a given PR number or URL, by triaging each unresolved thread, fixing what is warranted, pushing, and replying in each thread, when asked to address, fix or answer PR comments. Works in any Scenario repository.
中文简介正在准备中。当前显示原文说明。
Generate one short Conventional Commits squash message describing only the net diff between the PR's base and head, set its header as the PR title, validate it with the repository's commitlint and both copy its description to the clipboard and display it, when asked for a squash message. Works in any Scenario repository, whatever its default branch.
中文简介正在准备中。当前显示原文说明。
Refresh the current pull request description from its net changes, following the repository's PR template, when asked to update or write the PR summary. Works in any Scenario repository, whatever its default branch.
中文简介正在准备中。当前显示原文说明。
Use when connecting an AI agent to Scenario (scenario.com) through MCP, or when a task involves generating images, video, 3D, audio, sprites, textures, or game assets. Also when picking a Scenario model, running a LoRA, refining a generation prompt, uploading reference images, waiting on generation jobs, checking credits or quota, hitting Scenario auth, scope, or Forbidden errors, or setting up mcp.scenario.com in Claude Code, Cursor, VSCode, or another agent.
中文简介正在准备中。当前显示原文说明。
Use when generating or handling 3D assets through the Scenario MCP server, including text-to-3D or image-to-3D meshes, GPT-6 Astra 3D or Claude Opus 5.5 3D, GLB, FBX, OBJ, STL, or VOX files, PBR-textured or game-ready models, voxel models, multi-view reconstruction, editing a mesh by prompt, retexture, remesh, UV unwrap, auto-rigging a biped or quadruped, or retargeting an animation, previewing a mesh in the 3D viewer, or downloading a model for import into Unity, Unreal, Godot, or Blender.
中文简介正在准备中。当前显示原文说明。
Use when generating navigable 3D worlds or Gaussian splat scenes on Scenario via MCP: text-to-world, image-to-world, a 360 panorama or skybox turned into a walkable splat, multi-view photos or a video walkthrough reconstructed as a splat, or one object photo to splat. Keywords: World Labs Marble 1.1, 1.1 Plus, 1.0 Draft, Hunyuan World, HY World, image-to-skybox, skybox-to-splat, multiview-to-splat, TripoSplat, gaussian splatting, 3DGS, spz, world generation, explorable scene, VR.
中文简介正在准备中。当前显示原文说明。