makaron-cli 0.13.2 → 0.13.4

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "makaron-cli",
-  "version": "0.13.2",
+  "version": "0.13.3",
   "description": "AI image editing, video generation, music creation, and marketplace skill workflows via CLI. Agents can self-register, install skills, create projects, and produce creative media.",
   "author": {
     "name": "Makaron AI",

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "makaron-cli",
-  "version": "0.13.2",
+  "version": "0.13.3",
   "description": "AI image editing, video generation, music creation, and marketplace skill workflows via CLI. Agents can self-register, install skills, create projects, and produce creative media.",
   "displayName": "Makaron",
   "shortDescription": "AI image/video/music creation from the terminal",

package/README.md

@@ -96,12 +96,9 @@ npx makaron-cli chat --project <id> --json -b "<prompt>"
 # Auto-create project (with or without images)
 npx makaron-cli chat --project auto --image photo.jpg --json -b "make it cinematic"
 npx makaron-cli chat --project auto --image img1.jpg --image img2.jpg --json -b "combine these"
-
-# Choose each model role explicitly
-npx makaron-cli chat --project auto --agent-model deepseek-v4-pro --image-model qwen "design a product poster"
 ```
 
-Model flags are role-specific: `--agent-model` controls reasoning and tool use, `--image-model` controls image generation/editing, and `--video-model` controls video generation. `--agent-model` accepts `auto|gpt-5.6-terra|gpt-5.6-sol|gpt-5.6-luna|grok-4.5|deepseek-v4-pro`. `auto` currently uses GPT-5.6 Terra. `MAKARON_AGENT_MODEL` can set the default for automation; the command flag takes precedence. The legacy `--model` flag remains temporarily supported with a deprecation warning.
+`chat` always routes agent, image, and video models automatically. Never pass `--agent-model`, `--image-model`, `--video-model`, or the legacy `--model` flag to `chat`; the CLI rejects them before starting a run. Model flags remain available only on explicit low-level commands such as `edit` and `video create`.
 
 Returns immediately:
 ```json
@@ -118,7 +115,7 @@ Returns immediately:
 | Fix one moment in a video from a screenshot | `npx makaron-cli chat --project <id> --image screenshot.png "@4 this frame should be Paris; only fix this moment"` |
 | Cut or assemble video | `npx makaron-cli chat --project <id> --video clip.mp4 "cut out the dead air and keep the best 20 seconds"` |
 | Add music | `npx makaron-cli chat --project <id> "add calm piano background music"` |
-| Beat-sync video from audio | `npx makaron-cli chat --project auto --audio beat.mp3 --video-model seedance-fast --video-resolution 480p "make a beat-synced video"` |
+| Beat-sync video from audio | `npx makaron-cli chat --project auto --audio beat.mp3 "use Seedance Mini at 480p to make a beat-synced video"` |
 | Create motion design | `npx makaron-cli chat --project <id> "make an animated Instagram story with this image"` |
 
 ### Marketplace skills
@@ -219,13 +216,13 @@ Attach a short song, beat, or voice recording when the video should follow audio
 ```bash
 npx makaron-cli chat --project auto \
   --audio beat.mp3 \
-  --video-model seedance-fast \
-  --video-resolution 480p \
-  -b "make a 15s beat-synced video"
+  -b "use Seedance Mini at 480p to make a 15s beat-synced video"
 ```
 
 `--audio` accepts repeatable local files or public URLs. Local MP3/WAV files must be 2-15s and <=15MB; reference audio currently works with Seedance video generation.
 
+`chat` intentionally has no video model or resolution flags. State both in the chat message so the Agent selects a compatible provider and resolution together. Use `video create` only when you explicitly need direct provider controls.
+
 ### Fix one video moment from a screenshot
 
 When a video is mostly good but one moment needs a local fix, attach a screenshot of the problem frame and describe the correction in normal language:
@@ -307,8 +304,9 @@ npx makaron-cli video create --script "Shot 1 (5s): <<<image_1>>> ..." --image h
 npx makaron-cli video create --script "Shot 1 (5s): <<<image_1>>> slow cinematic push-in with native ambience" --image https://...jpg --duration 5 --video-model grok
 npx makaron-cli video create --script "Shot 1 (15s): <<<image_1>>> and <<<image_2>>> build a neon one-person studio" --image https://...jpg --image https://...webp --duration 15 --video-model seedance-mini --video-resolution 480p --aspect 9:16
 
-# 3b. Native SeeDance text-to-video (no image required)
+# 3b. Native SeeDance or MiniMax H3 text-to-video (no image required)
 npx makaron-cli video create --script "Shot 1 (5s): A neon one-person studio wakes at dawn" --duration 5 --video-model seedance-fast --aspect 16:9
+npx makaron-cli video create --script "Shot 1 (15s): A premium creative editor comes alive" --duration 15 --video-model minimax-h3 --video-resolution 2k --aspect 16:9
 
 # 3c. Edit a video from a local file or public URL
 npx makaron-cli video create --script "make it funny" --video input.mp4 --duration 5 --video-model seedance-fast
@@ -324,9 +322,11 @@ For project/timeline video editing, use:
 npx makaron-cli chat --project <id|auto> --video input.mp4 -b "make it funny"
 ```
 
-Options for `video create`: `--script "..."`, `--script-file <path>`, `--image <url>` (repeatable, up to 7), `--video <file|url>`, `--duration <seconds>`, `--aspect 9:16|16:9|1:1`, `--video-model seedance-fast|seedance-mini|seedance|kling|grok|google-omni`, `--video-resolution auto|480p|720p|1080p|4k`. Default model is `seedance-fast`. SeeDance accepts native text-to-video with no image and integer output duration 4-15s (default 5s); `seedance-mini` supports 480p/720p and is best for cheaper drafts/multi-size tests; Kling supports 5-15s; Grok 1.5 supports 1-15s single-image-to-video only; Gemini Omni supports 3-10s fast 720p image/video generation and editing with native generated audio, including up to 6 image references when no video reference is provided. For `--video-model grok`, forced `--aspect` is ignored to avoid xAI stretching the source image; pad/create the image at the target shape first or use another model.
+Options for `video create`: `--script "..."`, `--script-file <path>`, `--image <url>` (repeatable, up to the selected model limit), `--video <file|url>` and `--audio <file|url>` (repeatable where supported), `--duration <seconds>`, `--aspect 9:16|16:9|1:1`, `--video-model seedance-fast|seedance-mini|seedance|seedance-2.5|kling|grok|google-omni|minimax-h3`, `--video-resolution auto|480p|720p|768p|1080p|2k|4k`. Default model is `seedance-fast`. SeeDance accepts native text-to-video with no image and integer output duration 4-15s (default 5s); `seedance-mini` supports 480p/720p and is best for cheaper drafts/multi-size tests; MiniMax H3 accepts native text-to-video, 4-15s output, and up to 9 image, up to 3 video, and up to 3 audio references through Makaron Agent/chat. H3 defaults to public 2k output; 768p requires provider preview access plus `MINIMAX_H3_ENABLE_768P=true` on the Makaron server. Kling supports 5-15s; Grok 1.5 supports 1-15s single-image-to-video only; Gemini Omni supports 3-10s fast 720p image/video generation and editing with native generated audio, including up to 6 image references when no video reference is provided. For `--video-model grok`, forced `--aspect` is ignored to avoid xAI stretching the source image; pad/create the image at the target shape first or use another model.
+
+Seedance 2.5: use `--video-model seedance-2.5` for 4-30 second output at 480p/720p. `--image` accepts local files or URLs (up to 30), while repeatable `--video` and `--audio` accept up to 10 each. Use `--video-operation generate|edit|extend`, `--extend-direction forward|backward`, `--output-format mp4|mov`, `--web-search`, `--generated-audio` / `--no-generated-audio`, and `--relaxed-content-filter`. Edit/extend require a video reference. The Evolink route does not currently expose 4K output.
 
-Video edit model behavior: `--video-model kling --video` uses Kling base/direct edit internally; `--video-model seedance-fast --video`, `--video-model seedance-mini --video`, or `--video-model seedance --video` uses the SeeDance video-reference path and requires target <=15s, <=50MB, width/height 300-6000px, aspect ratio 0.4-2.5, and frame pixels 409,600-2,086,876. `--video-model google-omni --video` uses Gemini Omni direct video editing and accepts one reference video in Makaron. Output duration is clamped to 3-10s. Grok does not support video references.
+Video edit model behavior: `--video-model kling --video` uses Kling base/direct edit internally; `--video-model seedance-fast --video`, `--video-model seedance-mini --video`, or `--video-model seedance --video` uses the SeeDance video-reference path and requires target <=15s, <=50MB, width/height 300-6000px, aspect ratio 0.4-2.5, and frame pixels 409,600-2,086,876. `--video-model minimax-h3 --video` uses H3 feature/reference mode: up to 3 video references totaling <=15s, each <=50MB with width/height 256-5760px and aspect ratio 0.4-2.5. `--video-model google-omni --video` uses Gemini Omni direct video editing and accepts one reference video in Makaron. Output duration is clamped to 3-10s. Grok does not support video references.
 
 ### `music` — Music generation
 
@@ -439,7 +439,7 @@ send_message "All done!"
 ## Important Notes
 
 - One project = one conversation thread. All history is preserved.
-- One run at a time per project. New message interrupts previous run.
+- One active Agent Run at a time per project. A new message received while it is active is appended to that same Agent Run and processed at a durable work-unit boundary; it does not interrupt the execution or create a second owner for an in-progress Studio workflow.
 - Multi-image: `create --image a.jpg --image b.jpg` or `chat --image ref.jpg`.
 - Provider-generated videos can take 3-5 minutes; Grok is usually around 30-40 seconds; Gemini Omni is usually around 30-70 seconds plus Storage handoff. Remotion compositions should be converted with `materialize` / `responses get --materialize`, and timing should be read from `duration_seconds`, `render_seconds`, and `realtime_ratio`.
 - Music takes ~60 seconds. Appears in output when done.

package/bin/makaron.mjs

@@ -27,7 +27,6 @@ const APP_URL = process.env.MAKARON_APP_URL || DEFAULT_URL;
 const NPM_PACKAGE_NAME = 'makaron-cli';
 const UPDATE_CHECK_INTERVAL_MS = 24 * 60 * 60 * 1000;
 const UPDATE_CHECK_TIMEOUT_MS = 400;
-const AGENT_MODELS = ['auto', 'gpt-5.6-terra', 'gpt-5.6-sol', 'gpt-5.6-luna', 'grok-4.5', 'deepseek-v4-pro'];
 const AGENT_WAIT_TIMEOUT_SECONDS = Math.max(900, Number(process.env.MAKARON_AGENT_WAIT_TIMEOUT_SECONDS || 10_800));
 
 // Public anon key (safe to embed — only enables auth, not data access)
@@ -46,24 +45,20 @@ const MAX_AUDIO_REFERENCE_DURATION_TOLERANCE = 0.5;
 const MAX_AUDIO_REFERENCE_FILE_SIZE_MB = 15;
 const MAX_AUDIO_REFERENCE_FILE_SIZE = MAX_AUDIO_REFERENCE_FILE_SIZE_MB * 1024 * 1024;
 const MAX_VIDEO_FRAME_PIXELS = 2_086_876;
+const SEEDANCE25_MAX_VIDEO_REFERENCE_DURATION = 30;
+const SEEDANCE25_MAX_VIDEO_FRAME_PIXELS = 8_295_044;
 const SEEDANCE_MIN_VIDEO_FRAME_PIXELS = 409_600;
 const SEEDANCE_MIN_VIDEO_SIDE = 300;
 const SEEDANCE_MAX_VIDEO_SIDE = 6000;
 const SEEDANCE_MIN_VIDEO_ASPECT = 0.4;
 const SEEDANCE_MAX_VIDEO_ASPECT = 2.5;
+const MINIMAX_H3_MIN_VIDEO_SIDE = 256;
+const MINIMAX_H3_MAX_VIDEO_SIDE = 5760;
 
 function warnLegacyModelFlag(replacement) {
   process.stderr.write(`⚠️  --model is deprecated here; use ${replacement}.\n`);
 }
 
-function validateAgentModel(value) {
-  if (!AGENT_MODELS.includes(value)) {
-    process.stderr.write(`❌ Unknown agent model: ${value}\nChoose one of: ${AGENT_MODELS.join(', ')}\n`);
-    process.exit(1);
-  }
-  return value;
-}
-
 function getCliVersion() {
   try {
     const pkg = JSON.parse(fs.readFileSync(new URL('../package.json', import.meta.url), 'utf-8'));
@@ -291,15 +286,14 @@ Options:
   --video <file|url>        Attach a video to the project timeline. Repeatable.
   --audio <file|url>        Attach a song, beat, or voice reference. MP3/WAV, repeatable.
   --skill <id|label|name>   Use an installed skill or auto-install a matched marketplace skill.
-  --image-model <name>      Image model: gemini, gemini-lite, qwen, openai, pony, or wai.
-  --video-model <name>      Preferred video model: seedance-fast, seedance-mini, seedance, kling, grok, or google-omni.
-  --agent-model <name>      Agent model: auto, gpt-5.6-terra, gpt-5.6-sol, gpt-5.6-luna, grok-4.5, or deepseek-v4-pro.
-  --video-resolution <res>  Video resolution: auto, 480p, 720p, 1080p, or 4k.
   --background, -b          Submit and print a runId.
   --json                    Output structured JSON.
   --stream                  Legacy live SSE stream.
   --help, -h                Show this help.
 
+Model routing is automatic in chat. Do not pass --agent-model, --image-model,
+--video-model, or the legacy --model flag.
+
 What you can ask:
   Image edit
     makaron chat --project <id> --image photo.jpg "remove the person in the background"
@@ -323,7 +317,7 @@ What you can ask:
     makaron chat --project <id> "add calm piano background music"
 
   Reference audio / beat sync
-    makaron chat --project auto --audio beat.mp3 --video-model seedance-fast --video-resolution 480p "用这个音乐做卡点视频"
+    makaron chat --project auto --audio beat.mp3 "用 Seedance Mini 480p 做卡点视频"
     makaron chat --project <id> --audio https://example.com/beat.mp3 "add this as the soundtrack"
 
   Motion design
@@ -360,10 +354,6 @@ async function streamAgent(baseUrl, headers, projectId, prompt, opts = {}) {
       projectId,
       prompt,
       headless: true,
-      ...(opts.preferredModel ? { preferredModel: opts.preferredModel } : {}),
-      ...(opts.videoModel ? { videoModel: opts.videoModel } : {}),
-      ...(opts.videoResolution ? { videoResolution: opts.videoResolution } : {}),
-      ...(opts.agentModel && opts.agentModel !== 'auto' ? { agentModel: opts.agentModel } : {}),
       ...(opts.uploadedVideoCount ? { uploadedVideoCount: opts.uploadedVideoCount } : {}),
       ...(opts.turnMediaCount ? { turnMediaCount: opts.turnMediaCount } : {}),
     }),
@@ -472,10 +462,6 @@ async function streamAgent(baseUrl, headers, projectId, prompt, opts = {}) {
 
 async function submitRun(baseUrl, headers, projectId, prompt, opts = {}) {
   const body = { projectId, prompt };
-  if (opts.preferredModel) body.preferredModel = opts.preferredModel;
-  if (opts.agentModel && opts.agentModel !== 'auto') body.agentModel = opts.agentModel;
-  if (opts.videoModel) body.videoModel = opts.videoModel;
-  if (opts.videoResolution) body.videoResolution = opts.videoResolution;
   if (opts.currentSnapshotIndex != null) body.currentSnapshotIndex = opts.currentSnapshotIndex;
   if (opts.isNsfw) body.isNsfw = opts.isNsfw;
   if (opts.audioAttachments?.length) body.audioAttachments = opts.audioAttachments;
@@ -1383,7 +1369,8 @@ function getAudioMimeFromExt(ext) {
   return null;
 }
 
-function validateAudioReferenceFile(audioPath) {
+function validateAudioReferenceFile(audioPath, options = {}) {
+  const maxDuration = options.maxDuration ?? MAX_AUDIO_REFERENCE_DURATION;
   if (!fs.existsSync(audioPath)) {
     return { ok: false, error: `Audio file not found: ${audioPath}` };
   }
@@ -1406,8 +1393,8 @@ function validateAudioReferenceFile(audioPath) {
   if (duration < MIN_AUDIO_REFERENCE_DURATION) {
     return { ok: false, error: `Audio too short: ${formatSeconds(duration)}s (min ${MIN_AUDIO_REFERENCE_DURATION}s).` };
   }
-  if (duration > MAX_AUDIO_REFERENCE_DURATION + MAX_AUDIO_REFERENCE_DURATION_TOLERANCE) {
-    return { ok: false, error: `Audio too long: ${formatSeconds(duration)}s (max ${MAX_AUDIO_REFERENCE_DURATION}s, with ${MAX_AUDIO_REFERENCE_DURATION_TOLERANCE}s metadata tolerance).` };
+  if (duration > maxDuration + MAX_AUDIO_REFERENCE_DURATION_TOLERANCE) {
+    return { ok: false, error: `Audio too long: ${formatSeconds(duration)}s (max ${maxDuration}s, with ${MAX_AUDIO_REFERENCE_DURATION_TOLERANCE}s metadata tolerance).` };
   }
   return { ok: true, mime, meta: { duration, fileSizeBytes: stat.size } };
 }
@@ -1489,16 +1476,19 @@ function validateVideoFile(videoPath, options = {}) {
   const maxSide = options.maxSide ?? Infinity;
   const minAspect = options.minAspect ?? 0;
   const maxAspect = options.maxAspect ?? Infinity;
+  const allowedExtensions = options.allowedExtensions ?? ['mp4', 'mov', 'webm'];
+  const maxFramePixels = options.maxFramePixels ?? MAX_VIDEO_FRAME_PIXELS;
+  const maxFileSize = options.maxFileSize ?? MAX_VIDEO_UPLOAD_FILE_SIZE;
   if (!fs.existsSync(videoPath)) {
     return { ok: false, error: `Video file not found: ${videoPath}` };
   }
   const stat = fs.statSync(videoPath);
-  if (stat.size > MAX_VIDEO_UPLOAD_FILE_SIZE) {
-    return { ok: false, error: `Video too large: ${(stat.size / 1024 / 1024).toFixed(1)}MB (max ${MAX_VIDEO_UPLOAD_FILE_SIZE_MB}MB). The CLI uploads directly to Storage; use the frontend to transcode larger videos first.` };
+  if (stat.size > maxFileSize) {
+    return { ok: false, error: `Video too large: ${(stat.size / 1024 / 1024).toFixed(1)}MB (max ${(maxFileSize / 1024 / 1024).toFixed(0)}MB).` };
   }
   const ext = path.extname(videoPath).slice(1).toLowerCase();
-  if (!['mp4', 'mov', 'webm'].includes(ext)) {
-    return { ok: false, error: `Unsupported video format: .${ext}. Use MP4, MOV, or WebM.` };
+  if (!allowedExtensions.includes(ext)) {
+    return { ok: false, error: `Unsupported video format: .${ext}. Use ${allowedExtensions.map(value => value.toUpperCase()).join(' or ')}.` };
   }
   const meta = probeLocalVideo(videoPath);
   if (!meta) {
@@ -1507,8 +1497,8 @@ function validateVideoFile(videoPath, options = {}) {
   if (meta.duration > maxDuration + durationTolerance) {
     return { ok: false, error: `Video too long: ${formatSeconds(meta.duration)}s (max ${maxDuration}s, with ${durationTolerance}s metadata tolerance)` };
   }
-  if (meta.width * meta.height > MAX_VIDEO_FRAME_PIXELS) {
-    return { ok: false, error: `Video resolution too high: ${meta.width}x${meta.height} (${meta.width * meta.height} px). Max is <=1080p (${MAX_VIDEO_FRAME_PIXELS} px). Re-upload through the frontend to transcode, or export a smaller video.` };
+  if (meta.width * meta.height > maxFramePixels) {
+    return { ok: false, error: `Video resolution too high: ${meta.width}x${meta.height} (${meta.width * meta.height} px). Max is ${maxFramePixels} pixels.` };
   }
   const framePixels = meta.width * meta.height;
   const aspect = meta.width / meta.height;
@@ -1634,26 +1624,17 @@ Commands:
 
   admin                              Admin commands (skills, upload, set-admin)
 
-Model selection:
-  --agent-model <name>               Reasoning/tool model: auto, gpt-5.6-terra, gpt-5.6-sol,
-                                     gpt-5.6-luna, grok-4.5, or deepseek-v4-pro
-  --image-model <name>               Image model: gemini, gemini-lite, qwen, openai,
-                                     pony, or wai
-  --video-model <name>               Video model: seedance-fast, seedance-mini, seedance,
-                                     kling, grok, or google-omni
-
 Examples:
-  makaron chat --project auto --agent-model deepseek-v4-pro "plan a launch poster"
-  makaron chat --project <id> --agent-model gpt-5.6-terra --image-model qwen "make it cinematic"
-  makaron chat --project <id> --video-model seedance-fast "turn this into a short video"
+  makaron chat --project auto "plan a launch poster"
+  makaron chat --project <id> "make it cinematic"
+  makaron chat --project <id> "turn this into a short video"
 
 Run makaron <command> --help for command-specific options.
-The legacy --model flag is deprecated; use the role-specific flags above.
+Chat chooses agent, image, and video models automatically.
 
 Environment:
   MAKARON_API_KEY       API key (mk_live_xxx) — recommended for agents
   MAKARON_URL           API base (default: ${DEFAULT_URL})
-  MAKARON_AGENT_MODEL   Default Agent model; --agent-model takes precedence
 `);
 }
 
@@ -1744,8 +1725,6 @@ function printHelp(topic, subtopic) {
     else if (subtopic === 'install') console.log('Usage: makaron skills install <marketplace-id|label> [--json]');
     else console.log(`Skill commands:
   skills list --built-in              List all built-in Makaron skills and Studio Run recipes
-  skills list --built-in --openmontage
-                                      List OpenMontage-native adapters only
   skills list                         List marketplace skills
   skills search <query>               Search marketplace skills
   skills show <id|label> --built-in   Show a built-in skill
@@ -1771,13 +1750,14 @@ Use with chat:
     console.log('Usage: makaron analyze --video <file|url> ["question"]');
   } else if (topic === 'video') {
     if (subtopic === 'script') console.log('Usage: makaron video script --image <file> [--image <file>] [--lang en|zh] "direction"');
-    else if (subtopic === 'create') console.log('Usage: makaron video create --script "..." [--image <url> | --video <public-url>] [--duration 10] [--aspect 9:16] [--video-model seedance-fast|seedance-mini|seedance|kling|grok|google-omni] [--video-resolution auto|480p|720p|1080p|4k] [--keep-original-sound]');
+    else if (subtopic === 'create') console.log('Usage: makaron video create --script "..." [--image <url> ...] [--video <url> ...] [--audio <url> ...] [--duration 10] [--aspect 9:16] [--video-model seedance-fast|seedance-mini|seedance|seedance-2.5|kling|grok|google-omni|minimax-h3] [--operation generate|edit|extend] [--video-resolution auto|480p|720p|768p|1080p|2k|4k] [--keep-original-sound]');
     else if (subtopic === 'status') console.log('Usage: makaron video status <taskId> | --snapshot <snapshotId> [--wait]');
     else console.log(`Video commands:
   video script --image <file> [--image <file>] "direction"   Write video script
   video create --script "..." --video-model seedance-fast    Native text-to-video (no image required)
+  video create --script "..." --video-model minimax-h3 --video-resolution 2k    MiniMax H3 native 2K text-to-video
   video create --script "..." --image <url> [--duration 10]  Submit video task
-  video create --script "..." --video <public-url> [--video-model seedance-fast|seedance-mini|seedance|kling|google-omni]  Edit a video (standalone; Grok does not support video refs)
+  video create --script "..." --video <public-url> [--video-model seedance-fast|seedance-mini|seedance|seedance-2.5|kling|google-omni|minimax-h3]  Edit/reference a video (Grok does not support video refs)
   video status <taskId>                                      Check video status
   video status --snapshot <snapshotId> [--wait]              Check v2 video snapshot
 `);
@@ -1888,10 +1868,6 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
   let background = false;
   let jsonOutput = false;
   let activeSkill = undefined;
-  let videoModel = undefined;
-  let videoResolution = undefined;
-  let preferredModel = undefined;
-  let agentModel = process.env.MAKARON_AGENT_MODEL || 'auto';
   for (let i = 1; i < args.length; i++) {
     if (args[i] === '--project' && args[i + 1]) projectId = args[++i];
     else if (args[i] === '--image' && args[i + 1]) chatImages.push(args[++i]);
@@ -1902,13 +1878,17 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
     else if (args[i] === '--stream') useStream = true;
     else if (args[i] === '--background' || args[i] === '-b') background = true;
     else if (args[i] === '--json') jsonOutput = true;
-    else if (args[i] === '--video-model' && args[i + 1]) videoModel = args[++i];
-    else if (args[i] === '--video-resolution' && args[i + 1]) videoResolution = args[++i];
-    else if (args[i] === '--image-model' && args[i + 1]) preferredModel = args[++i];
-    else if (args[i] === '--agent-model' && args[i + 1]) agentModel = args[++i];
-    else if (args[i] === '--model' && args[i + 1]) {
-      warnLegacyModelFlag('--image-model');
-      preferredModel = args[++i];
+    else if (args[i] === '--video-resolution' || args[i].startsWith('--video-resolution=')) {
+      process.stderr.write('❌ makaron chat chooses video model and resolution together. Put the requested resolution in your chat message, for example: "use MiniMax H3 at 2K".\n');
+      process.exit(1);
+    }
+    else if (
+      ['--agent-model', '--image-model', '--video-model', '--model'].includes(args[i])
+      || ['--agent-model=', '--image-model=', '--video-model=', '--model='].some(prefix => args[i].startsWith(prefix))
+    ) {
+      const flag = args[i].split('=')[0];
+      process.stderr.write(`❌ makaron chat chooses agent, image, and video models automatically. Remove ${flag} and retry.\n`);
+      process.exit(1);
     }
     else promptParts.push(args[i]);
   }
@@ -1918,7 +1898,6 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
     console.error('Run: makaron chat --help');
     process.exit(1);
   }
-  agentModel = validateAgentModel(agentModel);
   const { headers, baseUrl } = getAuth();
   // Split images into URLs vs local files
   const imageUrlList = chatImages.filter(p => p.startsWith('http://') || p.startsWith('https://'));
@@ -2017,7 +1996,7 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
   const resolvedSkill = await resolveChatSkill(baseUrl, headers, activeSkill);
 
   // Upload videos to project timeline (via /api/projects/create with videoUrls)
-  let finalPrompt = resolvedSkill ? `[Active skill: ${resolvedSkill}]\n${prompt}` : prompt;
+  const finalPrompt = resolvedSkill ? `[Active skill: ${resolvedSkill}]\n${prompt}` : prompt;
   let audioAttachments = [];
   if (chatAudios.length > 0) {
     const audioImports = [];
@@ -2108,10 +2087,6 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
   if (useStream) {
     // Legacy SSE mode
     const { results } = await streamAgent(baseUrl, headers, projectId, finalPrompt, {
-      videoModel,
-      videoResolution,
-      preferredModel,
-      agentModel,
       uploadedVideoCount: uploadedTurnVideoCount,
       turnMediaCount: uploadedTurnMediaCount,
     });
@@ -2124,10 +2099,6 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
   } else {
     // Default: fire-and-forget + poll
     const { runId } = await submitRun(baseUrl, headers, projectId, finalPrompt, {
-      videoModel,
-      videoResolution,
-      preferredModel,
-      agentModel,
       audioAttachments,
       uploadedVideoCount: uploadedTurnVideoCount,
       turnMediaCount: uploadedTurnMediaCount,
@@ -2320,10 +2291,7 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
 
   if (sub === 'list') {
     const builtIn = args.includes('--built-in');
-    let skills = builtIn ? await fetchBuiltInSkills(baseUrl) : await fetchMarketplaceSkills(baseUrl);
-    if (builtIn && args.includes('--openmontage')) {
-      skills = skills.filter(skill => skill.sourceProject === 'openmontage');
-    }
+    const skills = builtIn ? await fetchBuiltInSkills(baseUrl) : await fetchMarketplaceSkills(baseUrl);
     if (jsonOutput) console.log(JSON.stringify({ skills }, null, 2));
     else if (builtIn) printBuiltInSkills(skills);
     else printMarketplaceSkills(skills);
@@ -2365,8 +2333,6 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
   } else {
     console.log(`Skill commands:
   skills list --built-in              List all built-in Makaron skills and Studio Run recipes
-  skills list --built-in --openmontage
-                                      List OpenMontage-native adapters only
   skills list                         List marketplace skills
   skills search <query>               Search marketplace skills
   skills show <id|label> --built-in   Show a built-in skill
@@ -2462,11 +2428,15 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
 
   } else if (sub === 'create') {
     const images = [];
+    const videos = [];
+    const audios = [];
     let script = '', duration = undefined, aspectRatio = undefined, videoModel = undefined, videoResolution = undefined, wait = false;
-    let video = null, keepOriginalSound = false;
+    let keepOriginalSound = false, videoOperation = undefined, extendDirection = undefined, outputFormat = undefined;
+    let generateAudio = undefined, contentFilter = undefined, webSearch = false;
     for (let i = 2; i < args.length; i++) {
       if (args[i] === '--image' && args[i + 1]) images.push(args[++i]);
-      else if (args[i] === '--video' && args[i + 1]) video = args[++i];
+      else if (args[i] === '--video' && args[i + 1]) videos.push(args[++i]);
+      else if (args[i] === '--audio' && args[i + 1]) audios.push(args[++i]);
       else if (args[i] === '--script' && args[i + 1]) script = args[++i];
       else if (args[i] === '--script-file' && args[i + 1]) script = fs.readFileSync(args[++i], 'utf-8');
       else if (args[i] === '--duration' && args[i + 1]) duration = Number(args[++i]);
@@ -2479,6 +2449,13 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
       else if (args[i] === '--video-resolution' && args[i + 1]) videoResolution = args[++i];
       else if (args[i] === '--resolution' && args[i + 1]) videoResolution = args[++i];
       else if (args[i] === '--keep-original-sound') keepOriginalSound = true;
+      else if (args[i] === '--video-operation' && args[i + 1]) videoOperation = args[++i];
+      else if (args[i] === '--extend-direction' && args[i + 1]) extendDirection = args[++i];
+      else if (args[i] === '--output-format' && args[i + 1]) outputFormat = args[++i];
+      else if (args[i] === '--no-generated-audio') generateAudio = false;
+      else if (args[i] === '--generated-audio') generateAudio = true;
+      else if (args[i] === '--relaxed-content-filter') contentFilter = false;
+      else if (args[i] === '--web-search') webSearch = true;
       else if (args[i] === '--project') {
         console.error('Usage: video create no longer supports --project. Use: makaron chat --project <id> --video <file|url> "your request"');
         process.exit(1);
@@ -2486,27 +2463,53 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
       else if (args[i] === '--wait') wait = true;
     }
     const selectedVideoModel = videoModel || 'seedance-fast';
-    const isSeedanceModel = selectedVideoModel === 'seedance-fast' || selectedVideoModel === 'seedance-mini' || selectedVideoModel === 'seedance';
-    if (!script || (!images.length && !video && !isSeedanceModel)) {
-      console.error('Usage: makaron video create --script "..." [--image <url> | --video <public-url>] [--duration 10] [--aspect 9:16] [--video-model seedance-fast|seedance-mini|seedance|kling|grok|google-omni] [--video-resolution auto|480p|720p|1080p|4k] [--keep-original-sound]');
+    const isSeedance25 = selectedVideoModel === 'seedance-2.5';
+    const isSeedanceModel = selectedVideoModel === 'seedance-fast' || selectedVideoModel === 'seedance-mini' || selectedVideoModel === 'seedance' || isSeedance25;
+    const isMinimaxH3 = selectedVideoModel === 'minimax-h3';
+    const supportsNativeTextToVideo = isSeedanceModel || isMinimaxH3;
+    if (!script || (!images.length && !videos.length && !audios.length && !supportsNativeTextToVideo)) {
+      console.error('Usage: makaron video create --script "..." [--image <url>] [--video <file|url>] [--audio <file|url>] [--duration 30] [--video-model seedance-2.5|minimax-h3]');
       process.exit(1);
     }
+    if (isSeedance25 && images.length > 30) { console.error('Seedance 2.5 supports at most 30 image references.'); process.exit(1); }
+    if (isSeedance25 && videos.length > 10) { console.error('Seedance 2.5 supports at most 10 video references.'); process.exit(1); }
+    if (isSeedance25 && audios.length > 10) { console.error('Seedance 2.5 supports at most 10 audio references.'); process.exit(1); }
+    if (isMinimaxH3 && images.length > 9) { console.error('MiniMax H3 supports at most 9 image references.'); process.exit(1); }
+    if (isMinimaxH3 && videos.length > 3) { console.error('MiniMax H3 supports at most 3 video references.'); process.exit(1); }
+    if (isMinimaxH3 && audios.length > 3) { console.error('MiniMax H3 supports at most 3 audio references.'); process.exit(1); }
+    if (videoOperation && !['generate', 'edit', 'extend'].includes(videoOperation)) { console.error('--video-operation must be generate, edit, or extend.'); process.exit(1); }
+    if (extendDirection && !['forward', 'backward'].includes(extendDirection)) { console.error('--extend-direction must be forward or backward.'); process.exit(1); }
+    if (outputFormat && !['mp4', 'mov'].includes(outputFormat)) { console.error('--output-format must be mp4 or mov.'); process.exit(1); }
 
     if (wait) {
       console.error('Usage: --wait is only supported for project timeline tasks. Use chat --project for project video generation, or poll the returned taskId with video status.');
       process.exit(1);
     }
 
-    let videoUrl = isHttpUrl(video) ? video : null;
-    let inputVideoMeta = null;
-    if (videoUrl) {
-      process.stderr.write(`📹 Assuming public video URL already matches provider reference limits. Seedance requires ≤${MAX_VIDEO_PROVIDER_REFERENCE_DURATION}s, ≤50MB, sides 300-6000px, frame pixels 409,600-${MAX_VIDEO_FRAME_PIXELS}; Kling requires ≤200MB and ≤2K; Google Omni accepts one reference video in Makaron; Grok does not support video references.\n`);
+    const providerMaxDuration = isSeedance25 ? SEEDANCE25_MAX_VIDEO_REFERENCE_DURATION : MAX_VIDEO_PROVIDER_REFERENCE_DURATION;
+    const providerMaxPixels = isMinimaxH3 ? Infinity : isSeedance25 ? SEEDANCE25_MAX_VIDEO_FRAME_PIXELS : MAX_VIDEO_FRAME_PIXELS;
+    const localImages = images.filter(image => !isHttpUrl(image));
+    if (localImages.length) {
+      const uploadedImages = await uploadImageFilesViaSignedUrl(baseUrl, headers, undefined, localImages);
+      const uploadedByPath = new Map(localImages.map((image, index) => [image, uploadedImages[index]]));
+      images = images.map(image => isHttpUrl(image) ? image : uploadedByPath.get(image)).filter(Boolean);
     }
-    if (video && !videoUrl) {
+    const videoUrls = [];
+    let inputVideoMeta = null;
+    for (const video of videos) {
+      if (isHttpUrl(video)) { videoUrls.push(video); continue; }
       const valid = validateVideoFile(video, {
-        maxDuration: MAX_VIDEO_PROVIDER_REFERENCE_DURATION,
+        maxDuration: providerMaxDuration,
         durationTolerance: MAX_VIDEO_PROVIDER_REFERENCE_DURATION_TOLERANCE,
-        ...(selectedVideoModel === 'seedance' || selectedVideoModel === 'seedance-fast' || selectedVideoModel === 'seedance-mini' ? {
+        maxFramePixels: providerMaxPixels,
+        maxFileSize: isSeedance25 ? 200 * 1024 * 1024 : MAX_VIDEO_UPLOAD_FILE_SIZE,
+        ...(selectedVideoModel === 'minimax-h3' ? {
+          allowedExtensions: ['mp4', 'mov'],
+          minSide: MINIMAX_H3_MIN_VIDEO_SIDE,
+          maxSide: MINIMAX_H3_MAX_VIDEO_SIDE,
+          minAspect: SEEDANCE_MIN_VIDEO_ASPECT,
+          maxAspect: SEEDANCE_MAX_VIDEO_ASPECT,
+        } : isSeedanceModel ? {
           minFramePixels: SEEDANCE_MIN_VIDEO_FRAME_PIXELS,
           minSide: SEEDANCE_MIN_VIDEO_SIDE,
           maxSide: SEEDANCE_MAX_VIDEO_SIDE,
@@ -2515,22 +2518,38 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
         } : {}),
       });
       if (!valid.ok) { console.error(`❌ ${valid.error}`); process.exit(1); }
-      inputVideoMeta = valid.meta;
+      inputVideoMeta ||= valid.meta;
       process.stderr.write(`📹 Uploading ${path.basename(video)} (${(fs.statSync(video).size/1024/1024).toFixed(1)}MB)...\n`);
-      videoUrl = await uploadFileViaSignedUrl(baseUrl, headers, undefined, video, valid.mime);
-      if (!videoUrl) process.exit(1);
+      const uploaded = await uploadFileViaSignedUrl(baseUrl, headers, undefined, video, valid.mime);
+      if (!uploaded) process.exit(1);
+      videoUrls.push(uploaded);
       process.stderr.write(`📹 Uploaded: ${path.basename(video)}\n`);
     }
+    const audioUrls = [];
+    for (const audio of audios) {
+      if (isHttpUrl(audio)) { audioUrls.push(audio); continue; }
+      const valid = validateAudioReferenceFile(audio, { maxDuration: isSeedance25 ? 30 : MAX_AUDIO_REFERENCE_DURATION });
+      if (!valid.ok) { console.error(`❌ ${valid.error}`); process.exit(1); }
+      process.stderr.write(`🎵 Uploading ${path.basename(audio)}...\n`);
+      const uploaded = await uploadFileViaSignedUrl(baseUrl, headers, undefined, audio, valid.mime, { uploadKind: 'audio' });
+      if (!uploaded) process.exit(1);
+      audioUrls.push(uploaded);
+    }
     // Standalone MCP tool (no project timeline write)
     process.stderr.write('🎬 Submitting video...\n');
-    const vArgs = videoUrl
-      ? { videoUrl, editPrompt: script, images, videoModel: selectedVideoModel, videoResolution, referType: (selectedVideoModel === 'seedance' || selectedVideoModel === 'seedance-fast' || selectedVideoModel === 'seedance-mini') ? 'feature' : 'base' }
-      : { script, images, videoModel: selectedVideoModel, videoResolution };
-    const effectiveDuration = duration || (inputVideoMeta?.duration ? Math.min(MAX_VIDEO_PROVIDER_REFERENCE_DURATION, Math.round(inputVideoMeta.duration)) : undefined);
+    const resolvedOperation = videoOperation || (isSeedance25 && videoUrls.length ? 'edit' : 'generate');
+    const vArgs = isSeedance25
+      ? { script, images, videoUrls, audioUrls, videoModel: selectedVideoModel, videoResolution, operation: resolvedOperation, extendDirection, outputFormat, generateAudio, contentFilter, webSearch }
+      : isMinimaxH3
+        ? { script, images, videoUrls, audioUrls, videoModel: selectedVideoModel, videoResolution }
+      : videoUrls[0]
+        ? { videoUrl: videoUrls[0], editPrompt: script, images, videoModel: selectedVideoModel, videoResolution, referType: isSeedanceModel ? 'feature' : 'base' }
+        : { script, images, videoModel: selectedVideoModel, videoResolution };
+    const effectiveDuration = resolvedOperation === 'edit' ? undefined : duration || (inputVideoMeta?.duration ? Math.min(providerMaxDuration, Math.round(inputVideoMeta.duration)) : undefined);
     if (effectiveDuration) vArgs.duration = effectiveDuration;
     if (aspectRatio) vArgs.aspectRatio = aspectRatio;
-    if (keepOriginalSound && videoUrl) vArgs.keepOriginalSound = true;
-    const result = await callMcpTool(baseUrl, headers, videoUrl ? 'makaron_edit_video' : 'makaron_create_video', vArgs);
+    if (keepOriginalSound && videoUrls.length && !isSeedance25) vArgs.keepOriginalSound = true;
+    const result = await callMcpTool(baseUrl, headers, videoUrls.length && !isSeedance25 && !isMinimaxH3 ? 'makaron_edit_video' : 'makaron_create_video', vArgs);
     const text = result?.content?.find(c => c.type === 'text')?.text;
     if (text) {
       console.log(text);
@@ -2574,8 +2593,9 @@ if (!command || command === '--help' || command === '-h' || command === 'help')
     console.log(`Video commands:
   video script --image <file> [--image <file>] "direction"   Write video script
   video create --script "..." --video-model seedance-fast    Native text-to-video (no image required)
+  video create --script "..." --video-model minimax-h3 --video-resolution 2k    MiniMax H3 native 2K text-to-video
   video create --script "..." --image <url> [--duration 10]  Submit video task
-  video create --script "..." --video <public-url> [--video-model seedance-fast|seedance-mini|seedance|kling|google-omni]  Edit a video (standalone; Grok does not support video refs)
+  video create --script "..." --video <public-url> [--video-model seedance-fast|seedance-mini|seedance|seedance-2.5|kling|google-omni|minimax-h3]  Edit/reference a video (Grok does not support video refs)
   video status <taskId>                                      Check video status
   video status --snapshot <snapshotId> [--wait]              Check v2 video snapshot
 `);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "makaron-cli",
-  "version": "0.13.2",
+  "version": "0.13.4",
   "description": "Talk to Makaron Agent from the terminal — create projects, edit images, generate videos",
   "type": "module",
   "scripts": {

package/skills/makaron/SKILL.md

@@ -83,12 +83,9 @@ npx makaron-cli chat --project <id> --json -b "<prompt>"
 # Auto-create project (with or without images)
 npx makaron-cli chat --project auto --image photo.jpg --json -b "make it cinematic"
 npx makaron-cli chat --project auto --image img1.jpg --image img2.jpg --json -b "combine these"
-
-# Choose each model role explicitly
-npx makaron-cli chat --project auto --agent-model deepseek-v4-pro --image-model qwen "design a product poster"
 ```
 
-Model flags are role-specific: `--agent-model` controls reasoning and tool use, `--image-model` controls image generation/editing, and `--video-model` controls video generation. `--agent-model` accepts `auto|gpt-5.6-terra|gpt-5.6-sol|gpt-5.6-luna|grok-4.5|deepseek-v4-pro`. `auto` currently uses GPT-5.6 Terra. `MAKARON_AGENT_MODEL` can set the default for automation; the command flag takes precedence. The legacy `--model` flag remains temporarily supported with a deprecation warning.
+`chat` always routes agent, image, and video models automatically. Never pass `--agent-model`, `--image-model`, `--video-model`, or the legacy `--model` flag to `chat`; the CLI rejects them before starting a run. Model flags remain available only on explicit low-level commands such as `edit` and `video create`.
 
 Returns immediately:
 ```json
@@ -105,7 +102,7 @@ Returns immediately:
 | Fix one moment in a video from a screenshot | `npx makaron-cli chat --project <id> --image screenshot.png "@4 this frame should be Paris; only fix this moment"` |
 | Cut or assemble video | `npx makaron-cli chat --project <id> --video clip.mp4 "cut out the dead air and keep the best 20 seconds"` |
 | Add music | `npx makaron-cli chat --project <id> "add calm piano background music"` |
-| Beat-sync video from audio | `npx makaron-cli chat --project auto --audio beat.mp3 --video-model seedance-fast --video-resolution 480p "make a beat-synced video"` |
+| Beat-sync video from audio | `npx makaron-cli chat --project auto --audio beat.mp3 "use Seedance Mini at 480p to make a beat-synced video"` |
 | Create motion design | `npx makaron-cli chat --project <id> "make an animated Instagram story with this image"` |
 
 ### Marketplace skills
@@ -179,13 +176,13 @@ Attach a short song, beat, or voice recording when the video should follow audio
 ```bash
 npx makaron-cli chat --project auto \
   --audio beat.mp3 \
-  --video-model seedance-fast \
-  --video-resolution 480p \
-  -b "make a 15s beat-synced video"
+  -b "use Seedance Mini at 480p to make a 15s beat-synced video"
 ```
 
 `--audio` accepts repeatable local files or public URLs. Local MP3/WAV files must be 2-15s and <=15MB; reference audio currently works with Seedance video generation.
 
+`chat` intentionally has no video model or resolution flags. State both in the chat message so the Agent selects a compatible provider and resolution together. Use `video create` only when you explicitly need direct provider controls.
+
 ### Fix one video moment from a screenshot
 
 When a video is mostly good but one moment needs a local fix, attach a screenshot of the problem frame and describe the correction in normal language:
@@ -266,8 +263,9 @@ npx makaron-cli analyze --video input.mp4 "describe the key actions and pacing"
 npx makaron-cli video create --script "Shot 1 (5s): <<<image_1>>> ..." --image https://...jpg --duration 5 --video-model kling
 npx makaron-cli video create --script "Shot 1 (15s): <<<image_1>>> and <<<image_2>>> build a neon one-person studio" --image https://...jpg --image https://...webp --duration 15 --video-model seedance-mini --video-resolution 480p --aspect 9:16
 
-# 3b. Native SeeDance text-to-video (no image required)
+# 3b. Native SeeDance or MiniMax H3 text-to-video (no image required)
 npx makaron-cli video create --script "Shot 1 (5s): A neon one-person studio wakes at dawn" --duration 5 --video-model seedance-fast --aspect 16:9
+npx makaron-cli video create --script "Shot 1 (15s): A premium creative editor comes alive" --duration 15 --video-model minimax-h3 --video-resolution 2k --aspect 16:9
 
 # 3c. Edit a video from a local file or public URL
 npx makaron-cli video create --script "make it funny" --video input.mp4 --duration 5 --video-model seedance-fast
@@ -283,9 +281,11 @@ npx makaron-cli video status <taskId>
 npx makaron-cli chat --project <id|auto> --video input.mp4 -b "make it funny"
 ```
 
-Options for `video create`: `--script "..."`, `--script-file <path>`, `--image <url>` (repeatable, up to 7), `--video <file|url>`, `--duration <seconds>`, `--aspect 9:16|16:9|1:1`, `--video-model seedance-fast|seedance-mini|seedance|kling|grok|google-omni`, `--video-resolution auto|480p|720p|1080p|4k`. Default model is `seedance-fast`. SeeDance accepts native text-to-video with no image and integer output duration 4-15s (default 5s); `seedance-mini` supports 480p/720p and is best for cheaper drafts/multi-size tests; Kling supports 5-15s; Grok 1.5 supports 1-15s single-image-to-video only; Gemini Omni supports 3-10s fast 720p image/video generation and editing with native generated audio, including up to 6 image references when no video reference is provided. For `--video-model grok`, forced `--aspect` is ignored to avoid xAI stretching the source image; pad/create the image at the target shape first or use another model.
+Options for `video create`: `--script "..."`, `--script-file <path>`, `--image <url>` (repeatable, up to the selected model limit), `--video <file|url>` and `--audio <file|url>` (repeatable where supported), `--duration <seconds>`, `--aspect 9:16|16:9|1:1`, `--video-model seedance-fast|seedance-mini|seedance|seedance-2.5|kling|grok|google-omni|minimax-h3`, `--video-resolution auto|480p|720p|768p|1080p|2k|4k`. Default model is `seedance-fast`. SeeDance accepts native text-to-video with no image and integer output duration 4-15s (default 5s); `seedance-mini` supports 480p/720p and is best for cheaper drafts/multi-size tests; MiniMax H3 accepts native text-to-video, 4-15s output, and up to 9 image, up to 3 video, and up to 3 audio references through Makaron Agent/chat. H3 defaults to public 2k output; 768p requires provider preview access plus `MINIMAX_H3_ENABLE_768P=true` on the Makaron server. Kling supports 5-15s; Grok 1.5 supports 1-15s single-image-to-video only; Gemini Omni supports 3-10s fast 720p image/video generation and editing with native generated audio, including up to 6 image references when no video reference is provided. For `--video-model grok`, forced `--aspect` is ignored to avoid xAI stretching the source image; pad/create the image at the target shape first or use another model.
+
+Seedance 2.5 uses `--video-model seedance-2.5` and supports 4-30s at 480p/720p, up to 30 images + 10 videos + 10 audios, repeatable local/URL references, `--video-operation generate|edit|extend`, `--extend-direction`, `--output-format mp4|mov`, and `--web-search`. The Evolink route does not currently expose 4K output.
 
-Video edit model behavior: `--video-model kling --video` uses Kling base/direct edit internally; `--video-model seedance-fast --video`, `--video-model seedance-mini --video`, or `--video-model seedance --video` uses the SeeDance video-reference path and requires target <=15s, <=50MB, width/height 300-6000px, aspect ratio 0.4-2.5, and frame pixels 409,600-2,086,876. `--video-model google-omni --video` uses Gemini Omni direct video editing and accepts one reference video in Makaron. Output duration is clamped to 3-10s. Grok does not support video references.
+Video edit model behavior: `--video-model kling --video` uses Kling base/direct edit internally; `--video-model seedance-fast --video`, `--video-model seedance-mini --video`, or `--video-model seedance --video` uses the SeeDance video-reference path and requires target <=15s, <=50MB, width/height 300-6000px, aspect ratio 0.4-2.5, and frame pixels 409,600-2,086,876. `--video-model minimax-h3 --video` uses H3 feature/reference mode: up to 3 video references totaling <=15s, each <=50MB with width/height 256-5760px and aspect ratio 0.4-2.5. `--video-model google-omni --video` uses Gemini Omni direct video editing and accepts one reference video in Makaron. Output duration is clamped to 3-10s. Grok does not support video references.
 
 ### `music` — Music generation
 
@@ -416,7 +416,7 @@ send_message "All done!"
 ## Important Notes
 
 - One project = one conversation thread. All history is preserved.
-- One run at a time per project. New message interrupts previous run.
+- One active Agent Run at a time per project. A new message received while it is active is appended to that same Agent Run and processed at a durable work-unit boundary; it does not interrupt the execution or create a second owner for an in-progress Studio workflow.
 - Multi-image: `create --image a.jpg --image b.jpg` or `chat --image ref.jpg`.
 - Provider-generated videos can take 3-5 minutes; Grok is usually around 30-40 seconds; Gemini Omni is usually around 30-70 seconds plus Storage handoff. Remotion compositions should be converted with `materialize` / `responses get --materialize`, and timing should be read from `duration_seconds`, `render_seconds`, and `realtime_ratio`.
 - Music takes ~60 seconds. Appears in output when done.