@kolbo/mcp 1.114.1 → 1.115.0

package/README.md

@@ -148,7 +148,7 @@ Without the optional skill, the config block alone already exposes every tool
 | `generate_creative_director` | One brief → N coordinated scenes (image or video) |
 | `generate_music` | Text (+ optional lyrics) → song. Style, title, negative tags, length, and Suno fine-controls (style weight, weirdness, audio weight, persona / singing voice) |
 | `generate_speech` | Text + voice → spoken audio. `eleven_model` (eleven_v4 / eleven_v3 / eleven_multilingual_v2) picks which ElevenLabs model speaks an ElevenLabs voice (stock or own clone; omit = the voice's default; same price; `turbo` needs v4). Full expressive/style control: Google/Gemini named voice-direction presets (`style_instructions_preset_id`: warm/dramatic/whisper/…) or free-form `style_instructions`, preset styles + emotions (DeepDub / MiniMax / Cartesia), speed, accent/`language`, and per-provider voice settings (ElevenLabs stability/similarity/style/speaker boost/seed, `enforce_language`, `output_format` mp3/wav, `text_normalization`; DeepDub accent/variance/tempo, MiniMax pitch/volume/intensity/timbre), plus `audio_effect` (phone/radio/megaphone/small_room/hall/distant) + `audio_effect_strength` for any voice at no extra credits. Status returns the same fields for reuse. |
-| `generate_sound` | Text → sound effect. Duration, prompt influence, and per-provider controls (Stable Audio guidance, Kie loop/tempo/key, Seed-Audio voice/speed/volume/pitch + reference audio/image) |
+| `generate_sound` | Text → sound effect. Duration, prompt influence, and per-provider controls (Stable Audio guidance, Kie loop/tempo/key, Seed-Audio voice/speed/volume/pitch + reference audio/image, output format, Mirelo SFX 2.0 negative prompt / variations / seed / bitrate) |
 | `generate_3d` | Text or reference images → 3D model (GLB/FBX/OBJ/USDZ) |
 | `transcribe_audio` | Audio/video URL or file → text + SRT subtitles. Language, speaker diarization, audio-event tagging, and SRT formatting (words/line, lines/subtitle, caption stretch) |
 | `separate_audio_stems` | Audio/video URL (or a Kolbo `generation_id`) → Dialogue / Music / Effects / without-dialogue (M&E) layers. Kolbo's own masking pipeline with a speech classifier on top, so a centred engine or ambience is not handed back mislabelled as dialogue. 5 credits, runs inline |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@kolbo/mcp",
-  "version": "1.114.1",
+  "version": "1.115.0",
   "description": "Kolbo AI MCP Server - Generate images, videos, music, speech, and sound effects from Claude Code",
   "main": "src/index.js",
   "bin": {

package/skill/GENERATED.md

@@ -1,6 +1,6 @@
 # AUTO-GENERATED — do not edit
 
-This tree is mirrored from kolbo-code@7434517, the single source of truth.
+This tree is mirrored from kolbo-code@f2ab5de, the single source of truth.
 Canonical source: packages/opencode/skills/kolbo/
 Distribution: .github/workflows/sync-skill-to-plugin.yml
 

package/skill/SKILL.md

@@ -1,19 +1,8 @@
 ---
-version: 0.9.24
+version: 0.9.25
 name: kolbo
 description: |
-  Generate, edit, analyze, and direct creative media through Kolbo AI: images,
-  video (Seedance, Veo, Kling, Hailuo), music, speech, sound, 3D, transcription,
-  Visual DNA, Creative Director batches, marketing assets, HTML artifacts, and
-  AI Docs, Flow node workflows and session/system prompts, and approved Blender scene control. Use for sophisticated AI
-  filmmaking as well as individual media:
-  scripts, production bibles, recurring characters and locations, acting,
-  dialogue, music performance, blocking, physics, multi-shot continuity,
-  connected scenes, prompt audits, feature-length production planning, end-to-end vertical micro-drama series (Micro-Drama Studio), and
-  direct Blender scene building through the connected Kolbo Blender plugin.
-
-  NOT for: unrelated standalone video editing / FFmpeg (use video-production), motion graphics
-  (use remotion-best-practices), code editing, or general chat.
+  Generate, edit, analyze, and direct creative media through Kolbo AI: images, video (Seedance, Veo, Kling, Hailuo), music, speech, sound, 3D, transcription, Visual DNA, Creative Director batches, marketing assets, HTML artifacts, and AI Docs, Flow node workflows and session/system prompts, and approved Blender scene control. Use for sophisticated AI filmmaking as well as individual media: scripts, production bibles, recurring characters and locations, acting, dialogue, music performance, blocking, physics, multi-shot continuity, connected scenes, prompt audits, feature-length production planning, end-to-end vertical micro-drama series (Micro-Drama Studio), and direct Blender scene building through the connected Kolbo Blender plugin. NOT for: unrelated standalone video editing / FFmpeg (use video-production), motion graphics (use remotion-best-practices), code editing, or general chat.
 argument-hint: "[prompt-or-command] [--model <name>] [--image <path>] [--video <path>]"
 allowed-tools: Bash, Read, Write, Edit
 ---
@@ -280,7 +269,7 @@ Dependencies accumulate: narrative Elements work requires **Kolbo + filmmaking +
 
 ## ⚠️ Seedance / Elements prompt contract (HARD RULE)
 
-`generate_elements`, Seedance 2 / 2.5, WAN, MiniMax H3, Gemini, and any Visual DNA video share **one** compile shape — the Locked Intro in `references/models/seedance.md`. Load `elements-prompting` first (craft, `@Image N` mapping, eight elements), then compile:
+Mode wins over model family. **Frames** (`generate_video_from_image`) and **First/Last Frame** (`generate_first_last_frame`) use ordinary motion/camera/transition prose grounded in the supplied composition, including on Seedance 2.5. Do not add Elements CAST/LOCATION blocks, invent DNA names, or enable cuts because a template says so. An ordinary uploaded frame is `@Image1`, not a DNA; an end frame is `@Image2`. For **Elements / text-driven multi-shot** work, `generate_elements`, Seedance 2 / 2.5, WAN, MiniMax H3, Gemini, and reference-composed video share the Locked Intro in `references/models/seedance.md`. Load `elements-prompting` first (craft, exact `@ImageN` mapping, eight elements), then compile:
 
 `N connected cinematic shots, Xs total, AR, Multishot ON` → `Total: Xs / N shots / AR` → `[GLOBAL LOOK – LOCKED, APPLIES TO EVERY SHOT]` → `[CAST – IDENTICAL IN EVERY SHOT]` (each person is `@DNAName`) → `[LOCATION]` → LOCATION MAP / CONTINUITY / PHYSICS → `SHOT N — 0:00–0:02 — …` (ranges sum to Xs) → closing `Total: Xs / N shots / AR`. Pass MCP `duration: X` matching that Total. Omitting Total / Multishot is a failed compile — same contract as the Kolbo help widget.
 
@@ -341,6 +330,16 @@ A user-named tool — in any language — overrides every other rule. Recognized
 
 Model types for `list_models`: `text_to_img`, `image_editing`, `text_to_video`, `img_to_video`, `draw_to_video`, `video_to_video`, `elements`, `firstlastgenerations`, `lipsync-image`, `lipsync-video`, `music_gen`, `text_to_speech`, `text_to_sound`, `stt`, `text`, `3d_text_to_model`, `3d_image_to_model`, `3d_multi_image_to_model`, `3d_world`.
 
+### Practical iteration and prompt diagnosis
+
+- Start from the final deliverable. For a video, intermediate stills need only serve their intended role. Elements references can establish identity, wardrobe, props, environment or style without fixing every camera angle. Frames are literal compositions to animate. Do not make the user perfect every intermediate still when the final video is what matters.
+- Keep the original prompt, original references, model and settings when available. Repeatedly editing an imperfect derivative can compound drift. When a major viewpoint or composition edit keeps failing, compare a different edit wording with a fresh generation from the original reference. Preserve an explicit request to edit rather than silently switching routes.
+- An image prompt describes one still moment. Remove video timelines from an image request and write the desired still. A video prompt describes visible motion, camera trajectory, timing and sound. For an unbroken take, write timed actions within one shot; do not turn intermediate poses into cuts or dissolves.
+- Use economical trials only within the authorized scope and budget. A couple of samples can distinguish random variation from a repeated failure; a few genuinely different prompt approaches can test different mechanisms. These are options, not required extra spending. When comparing models, keep the prompt, references/order, duration and ratio comparable. Ratio affects composition and behavior; preserve the intended output format.
+- Judge the actual result, not whether an LLM praised the prompt. Separate reference binding/delivery failures from a poor source, an unsuitable method, model behavior and mixed conversation context. When an unchanged method fails repeatedly, change the tested hypothesis rather than keep spending on it. Do not promise that a prompt rewrite guarantees success.
+- Use exact numbered upload tokens in attachment order. A title, filename or invented subject name never becomes a registered DNA. Retain a supplied/selected DNA name verbatim, including Hebrew; never invent a DNA to fill a template's CAST block.
+- Keep unrelated goals in separate sessions. When context becomes mixed, preserve a compact checkpoint with the accepted brief, source references, working prompts/settings and observed failures. Offer a supported fork/new session; do not claim to have moved or forked a session without a successful tool result.
+
 ## ⚠️ Generation lifecycle — source of truth, waiting, failures (HARD RULE — read this)
 
 **How calls work:** each generation tool blocks until the job is fully complete. Images: seconds. Video: minutes. Multiple tool calls in one response run concurrently. On hosts with live widgets the tool instead returns `submitted` (or `_timed_out`) instantly — the card updates on its own.

package/skill/VERSION

@@ -1 +1 @@
-0.9.24
+0.9.25

package/skill/references/models/seedance.md

@@ -7,10 +7,12 @@
 
 Load this file when the user wants a **Seedance 2 / Seedance 2.0** (ByteDance) video. For any other video model, see `models/veo.md`, `models/prompt-copilot.md`, or generic video rules in `SKILL.md`.
 
-**Kolbo MCP routing:** Seedance is a video model — call `generate_video` (text-to-video) or `generate_elements` (when video references / Visual DNA / first-last frames are involved). Run `list_models({ type: "text_to_video" })` and pick a Seedance variant by name.
+**Kolbo MCP routing:** the selected inputs/mode choose the tool, not the Seedance name. Use `generate_video` for text-to-video, `generate_elements` for reference-composed scenes, `generate_video_from_image` to animate an opening still, and `generate_first_last_frame` for supplied opening/landing frames. Discover a matching variant and its constraints with the corresponding model type. First/last frames do not become Elements because they contain a recurring character.
 
 **Elements uses this same file.** `generate_elements` is not a second prompt language. Do not write `SCENE CONTEXT` / `OPTICS` / `ACTION` department packs for Elements or Seedance — those are filmmaking audit contracts, not the generation compile shape.
 
+**Frames format takes precedence:** on image-to-video or first/last-frame, write ordinary motion, camera and transition prose grounded in the supplied composition. Default to one continuous take; timed action beats are allowed. Omit Locked Intro/CAST/LOCATION blocks and multi-shot headers unless the user's selected route actually requires them. Use `@Image1` for the opening still and `@Image2` for a supplied landing still; never invent a DNA name or copy a placeholder identifier. Existing supplied DNA tags retain their exact spelling.
+
 ## Creative direction takes precedence
 
 The current user brief overrides template defaults and illustrative examples. Keep the two-layer organization, but include only relevant locks. State concrete camera trajectory and visible action prominently; optics numbers, equipment names, repetition and word counts are not guarantees of fidelity. Preserve a continuous-shot exception even when other scenes are multishot. For one shot use `Single continuous shot`, `Total: Xs / 1 shot / AR`, one SHOT heading and `multi_shots: false`; for multiple shots use `Multishot ON` and matching counts. AR comes from the brief, never a copied example. Seedance does not speak Hebrew — use Latin transliteration in quotes or route native Hebrew to Gemini Omni. Narration reserved for post does not belong in the generation prompt.
@@ -128,7 +130,7 @@ FORBIDDEN: omitting Total / Multishot; "same character throughout" as the only l
 
 ### 6. Reference-Anchored Cinematic Sequence (multi-character / named references — highest-fidelity format)
 
-Use whenever the user gives named characters, Visual DNAs, or multiple reference images (`@Image1`, `@Image2`, …). **This is always an Elements-mode prompt** (`generate_elements`). Structure:
+Use for **reference-composed Elements scenes** with named characters, Visual DNAs, or multiple reference assets (`@Image1`, `@Image2`, …). This is an Elements-mode prompt (`generate_elements`), not a replacement for explicitly selected Frames/first-last mode. Structure:
 
 1. **Locked Intro FIRST** (GLOBAL LOOK / CAST / LOCATION). CAST is one line per person: `@Exact_DNA_name` or `Name @ImageN — wardrobe, position, what they carry`. End CAST with "All must match their character references exactly." Never "the man on the LEFT" without an `@tag`.
 2. **REFERENCE CONSISTENCY block** — map every reference and pin what must NOT change: `@Zohar` / `Reference Image 1 is <X>. Preserve exact face, hair, anatomy, wardrobe, colors, props.`

package/skill/references/models/seedance25.md

@@ -7,7 +7,7 @@
 
 Load this file when the user wants a **Seedance 2.5** video (they said "2.5" / "25", or they need longer than 15s, more than ~10 shots, or a large cast of references). Also load `models/seedance.md` for the shared craft layer.
 
-**Kolbo MCP routing:** `generate_video` or `generate_elements` (refs / Visual DNA / first-last). Run `list_models({ type: "text_to_video" })` and pick the Seedance 2.5 variant by name.
+**Kolbo MCP routing:** preserve the selected mode. `generate_video` is text-to-video; `generate_elements` composes reference assets; `generate_video_from_image` animates an opening still; `generate_first_last_frame` connects supplied opening/landing frames. Discover the compatible Seedance 2.5 variant using that tool's model type. Frames use ordinary motion/transition prose, not the Elements Locked Intro. An uploaded frame is `@Image1` (landing frame `@Image2`), never an invented DNA name.
 
 **Audio:** Seedance 2.5 emits real synced audio. `list_models` shows `sound_generation_type: none` only because there is no in-app toggle (`sound_baked_in: true`) — it does NOT mean the model is silent, and it is never a reason to reach for TTS. Quoted dialogue is PERFORMED (synced voices, lip movement, room tone) alongside the SFX named in AUDIO, so scene dialogue never goes through `generate_speech` or `generate_lipsync`; write the lines in quotes inside their shot beats.
 

package/skill/references/workflows/prompt-structure.md

@@ -48,6 +48,8 @@ Rules that make it work rather than just look organised:
 
 ## 2. The reference contract — how an edit stops drifting
 
+For iterative work, retain the original source prompt and reference as the baseline. An edit request should preserve only the properties that must survive. If a major camera/viewpoint change repeatedly resists the same edit method, offer a materially different prompt approach or a fresh generation grounded in the original reference; do not keep chaining degraded outputs. Preserve an explicitly selected edit route unless the user accepts changing it. Diagnose with comparable, budgeted samples, not unbounded retries. Image prompts describe one still moment; video timing/action belongs in the video prompt.
+
 This is the highest-value pattern in the file. "Keep everything else the same" is too
 vague to hold, and it is why a chain of edits quietly turns a person into someone else.
 An edit prompt opens with an explicit contract:

package/src/tools/generate.js

@@ -1234,7 +1234,7 @@ function registerGenerateTools(server, client, options = {}) {
   // ─── generate_sound ────────────────────────────────────────
   server.tool(
     'generate_sound',
-    'Generate sound effects (not music, not speech) from a text description using Kolbo AI. Use this for ambient sounds, foley, impacts, atmospheres, UI sounds, etc. For music use generate_music; for voice use generate_speech. Beyond the core prompt/duration, per-provider controls are available (Stable Audio guidance, Kie loop/tempo/key, Seed-Audio voice/speed/volume/pitch + reference audio/image); the engine ignores controls that do not apply to the chosen model. Returns the final audio URL when complete.',
+    'Generate sound effects (not music, not speech) from a text description using Kolbo AI. Use this for ambient sounds, foley, impacts, atmospheres, UI sounds, etc. For music use generate_music; for voice use generate_speech. Beyond the core prompt/duration, per-provider controls are available (Stable Audio guidance, Kie loop/tempo/key, Seed-Audio voice/speed/volume/pitch + reference audio/image, output audio_format, and Mirelo SFX 2.0 negative_prompt / num_samples / seed / audio_bitrate_kbps); the engine ignores controls that do not apply to the chosen model. Mirelo SFX 2.0 (model "mirelo-ai/sfx2.0/text-to-audio") makes up to 10 minutes of layered sound design and up to 4 variations per call (each variation is billed). Returns the final audio URL(s) when complete.',
     {
       prompt: z.string().describe('Text description of the sound effect (e.g., "thunder clap with rain", "door creaking open", "futuristic UI beep")'),
       model: z.string().optional().describe('Model identifier. Use list_models type="text_to_sound" to see options. Default: elevenlabs-sound-effects-v1'),
@@ -1253,21 +1253,30 @@ function registerGenerateTools(server, client, options = {}) {
       seed_pitch: z.number().optional().describe('FAL Seed-Audio: pitch shift in semitones.'),
       seed_reference_audio_urls: z.array(z.string()).optional().describe('FAL Seed-Audio: up to 3 reference audio URLs to condition the sound.'),
       seed_reference_image_url: z.string().optional().describe('FAL Seed-Audio: a reference image URL to condition the sound.'),
+      // ── Output format (models whose list_models entry has output_audio_formats) ──
+      audio_format: z.enum(['wav', 'mp3', 'aac', 'flac']).optional().describe('Output audio container/codec, for models that advertise output_audio_formats in list_models (Sonilo, Mirelo SFX 2.0).'),
+      // ── Mirelo SFX 2.0 (list_models: max_samples / supports_negative_prompt / supports_seed / supports_audio_bitrate) ──
+      negative_prompt: z.string().max(1000).optional().describe('Mirelo SFX 2.0: sounds to keep OUT of the result, e.g. "music, speech".'),
+      num_samples: z.number().int().min(1).max(4).optional().describe('Mirelo SFX 2.0: number of variations (1-4, default 1). Each variation is billed; all URLs are returned.'),
+      seed: z.number().int().min(-1).optional().describe('Mirelo SFX 2.0: seed for repeatable results (-1 or omit = random).'),
+      audio_bitrate_kbps: z.number().int().optional().describe('Mirelo SFX 2.0: bitrate for mp3 (32,40,48,56,64,80,96,112,128,160,192,224,256,320) or aac (32-256). Not allowed with wav/flac.'),
       project_id: projectIdField,
       session_id: sessionIdField
     },
-    async ({ prompt, model, duration, prompt_influence, cfg_strength, sound_loop, sound_tempo, sound_key, seed_voice, seed_speed, seed_volume, seed_pitch, seed_reference_audio_urls, seed_reference_image_url, project_id, session_id }) => {
+    async ({ prompt, model, duration, prompt_influence, cfg_strength, sound_loop, sound_tempo, sound_key, seed_voice, seed_speed, seed_volume, seed_pitch, seed_reference_audio_urls, seed_reference_image_url, audio_format, negative_prompt, num_samples, seed, audio_bitrate_kbps, project_id, session_id }) => {
       model = await canonicalModelId(client, model, 'text_to_sound'); // lenient id resolution ("z-image" → "z-image/turbo")
       const gen = await client.post('/v1/generate/sound', {
         prompt, model, duration, prompt_influence,
         cfg_strength, sound_loop, sound_tempo, sound_key,
         seed_voice, seed_speed, seed_volume, seed_pitch,
-        seed_reference_audio_urls, seed_reference_image_url, project_id, session_id
+        seed_reference_audio_urls, seed_reference_image_url,
+        audio_format, negative_prompt, num_samples, seed, audio_bitrate_kbps,
+        project_id, session_id
       });
 
       if (returnsImmediately()) return submittedResult({
         tool: 'generate_sound', kind: 'audio', gen, client, model, prompt,
-        settings: { duration },
+        settings: { duration, ...(num_samples > 1 && { variations: num_samples }) },
         reference_images: seed_reference_image_url ? [seed_reference_image_url] : []
       });
 
@@ -1280,7 +1289,7 @@ function registerGenerateTools(server, client, options = {}) {
 
       return uiCompleted({
         tool: 'generate_sound', kind: 'audio', gen, client, model, prompt,
-        settings: { duration },
+        settings: { duration, ...(num_samples > 1 && { variations: num_samples }) },
         reference_images: seed_reference_image_url ? [seed_reference_image_url] : [],
         urls: result.result.urls,
         playback_urls: result.result.playback_urls,
@@ -2391,6 +2400,10 @@ function registerGenerateTools(server, client, options = {}) {
   const editDetails = (op, a) => {
     const d = [];
     const has = (v) => v !== undefined && v !== null && v !== '';
+    if (op === 'generate_audio') {
+      if (a.num_samples > 1) d.push(`${a.num_samples} variations`);
+      if (a.preserve_speech === false) d.push('no dialogue');
+    }
     if (op === 'upscale' || op === 'split_upscale') {
       if (has(a.scale)) d.push(`${a.scale}×`);
       if (a.face_enhancement === false) d.push('no face enhance');
@@ -2678,7 +2691,7 @@ function registerGenerateTools(server, client, options = {}) {
         '"draft_quote" — check a saved draft’s supported final resolutions, expiry and exact credit cost without starting a render. Use this before draft_enhance.',
         '"upscale" — boost to 4K/2K resolution (use `scale` for factor, `resolution` for target, `target_fps` for frame rate). On model "blackforestlabs/flux-video-upscale" (Flux Video Upscale): `scale` is 1.5-3x (no named `resolution` target — billed on the OUTPUT tier it lands in), `mode` is "precise" (source-faithful, default) or "creative" (reimagines detail — a pricier tier), and `prompt` optionally guides the creative-mode enhancement. On model "bytedance-upscaler/upscale/video" (Bytedance Upscaler): `resolution` is "1080p" | "2k" | "4k" | "6k" | "8k" (OR `scale_ratio` 1.1-10, output ≤4K), plus `target_fps` (24-120), `enhancement_tier`, `enhancement_preset`, `fidelity`, `bit_depth` — see each field. Price = resolution tier (1080p 1 · 2k 2 · 4k 4 · 6k 8 · 8k 16 credits/s) × every started 30 fps (60 = 2×, 120 = 4×) × tier (fast 0.5×, pro 10×). Source-length caps: 6K 120s, 8K 60s, Pro 30s, otherwise 300s.',
         '"reframe" — change aspect ratio (requires `aspect_ratio`; use `grid_position_x`/`grid_position_y` to control where the original sits).',
-        '"generate_audio" — add AI-generated audio from `prompt`. Optionally split into `sound_effect_prompt` and `background_music_prompt`. Set `original_sound=true` to keep original audio alongside.',
+        '"generate_audio" — add AI-generated audio from `prompt`. Optionally split into `sound_effect_prompt` and `background_music_prompt`. Set `original_sound=true` to keep original audio alongside. On model "mirelo-ai/sfx2.0/video-to-video" (Mirelo SFX 2.0) the prompt is optional (it watches the clip and builds a full synced sound design, up to 10 minutes): `preserve_speech` keeps the dialogue (default true), `negative_prompt` keeps sounds out, `num_samples` returns 1-4 variations (each billed), `seed` repeats a result, and `duration` makes a shorter soundtrack by TRIMMING the video (omit for the whole clip).',
         '"remove_watermark" — AI-powered watermark removal.',
         '"face_swap" — replace the face in the video with the face from `image_url` (required).',
         '"extend" — lengthen the video at `mode` ("start" or "end") by `duration` seconds. Optional `prompt` and `context` guide the generated content.',
@@ -2742,6 +2755,15 @@ function registerGenerateTools(server, client, options = {}) {
         prompt: z.string()
       })).optional()
         .describe('Optional contiguous Sonilo sound-design ranges. The first start must be 0, each end must equal the next start, and the final end must not exceed the video duration. Omit to let Sonilo detect scenes automatically.'),
+      // Mirelo SFX 2.0 (list_models: max_samples / supports_negative_prompt / supports_seed / supports_preserve_speech)
+      negative_prompt: z.string().max(1000).optional()
+        .describe('Mirelo SFX 2.0 (operation="generate_audio"): sounds to keep out, as a short comma-separated list, e.g. "music, crowd noise".'),
+      preserve_speech: z.boolean().optional()
+        .describe('Mirelo SFX 2.0 (operation="generate_audio"): true (default) keeps the dialogue and builds the soundtrack around it; false makes sound effects only and suppresses speech.'),
+      num_samples: z.number().int().min(1).max(4).optional()
+        .describe('Mirelo SFX 2.0 (operation="generate_audio"): number of variations, 1-4 (default 1). Each is billed; all result URLs are returned.'),
+      seed: z.number().int().min(-1).optional()
+        .describe('Mirelo SFX 2.0 (operation="generate_audio"): seed for repeatable results (-1 or omit = random).'),
 
       // ── face_swap ──────────────────────────────────────────
       image_url: z.string().optional()
@@ -2749,7 +2771,7 @@ function registerGenerateTools(server, client, options = {}) {
 
       // ── extend ─────────────────────────────────────────────
       duration: z.number().optional()
-        .describe('Seconds of content to generate. Used with "extend" (required) and "retake" (optional). Typical range: 1–20.'),
+        .describe('Seconds of content to generate. Used with "extend" (required) and "retake" (optional). Typical range: 1–20. With "generate_audio" on Mirelo SFX 2.0: seconds of sound (1-600) — the returned video is TRIMMED to this length; omit for the whole clip.'),
       mode: z.string().optional()
         .describe('Extension direction: "start" or "end". Used with "extend". For "retake": replacement mode (default: "replace_audio_and_video"). For "upscale" on model "blackforestlabs/flux-video-upscale": "precise" (source-faithful, default) or "creative" (reimagines detail — pricier). Default: "end".'),
       context: z.string().optional()
@@ -2790,6 +2812,7 @@ function registerGenerateTools(server, client, options = {}) {
       grid_position_x, grid_position_y,
       sound_effect_prompt, background_music_prompt, original_sound, cfg_strength,
       audio_format, segments,
+      negative_prompt, preserve_speech, num_samples, seed,
       refine_edges, subject_is_person,
       text_prompt, context,
       mask_video_url, object_prompt, video_strength,
@@ -2806,7 +2829,7 @@ function registerGenerateTools(server, client, options = {}) {
       // Validation
       if (operation === 'magic_edit'    && !prompt)      throw new Error('prompt is required for magic_edit');
       if (operation === 'generate_audio' && !prompt && !sound_effect_prompt && !background_music_prompt
-          && !(typeof model === 'string' && model.includes('sonilo'))) {
+          && !(typeof model === 'string' && (model.includes('sonilo') || model.includes('mirelo')))) {
         throw new Error('prompt is required for generate_audio unless dedicated sound/music prompts are provided or the selected model supports automatic captioning');
       }
       if (operation === 'reframe'       && !aspect_ratio)throw new Error('aspect_ratio is required for reframe');
@@ -2824,6 +2847,7 @@ function registerGenerateTools(server, client, options = {}) {
         grid_position_x, grid_position_y,
         sound_effect_prompt, background_music_prompt, original_sound, cfg_strength,
         audio_format, segments,
+        negative_prompt, preserve_speech, num_samples, seed,
         refine_edges, subject_is_person,
         text_prompt, context,
         mask_video_url, object_prompt, video_strength,
@@ -2840,7 +2864,7 @@ function registerGenerateTools(server, client, options = {}) {
       if (operation === 'draft_quote') return { content: [{ type: 'text', text: JSON.stringify(gen, null, 2) }] };
       const cardSettings = {
         mode: operation, duration, aspect_ratio, resolution,
-        details: editDetails(operation, { scale, enhancement_model, target_fps, mode, enhancement_tier, enhancement_preset, bit_depth, scale_ratio }),
+        details: editDetails(operation, { scale, enhancement_model, target_fps, mode, enhancement_tier, enhancement_preset, bit_depth, scale_ratio, num_samples, preserve_speech }),
       };
       if (returnsImmediately()) return submittedResult({
         tool: 'edit_video', kind: 'video', gen, client, model,