ottoport 1.9.0 → 1.11.0

package/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "ottoport",
   "description": "One API for every model. Call chat, image, video, speech and music models through the OttoPort gateway — as MCP tools, slash commands, or the bundled CLI.",
-  "version": "1.9.0",
+  "version": "1.11.0",
   "author": {
     "name": "LITBOX LLC",
     "email": "support@ottoport.ai"

package/README.md

@@ -84,13 +84,14 @@ it needs Node 20+ on the PATH of whatever starts that host.
 
 ### Modes and resolution
 
-Video takes four modes — text-to-video, image-to-video (`image_url`),
-first-and-last-frame (`image_url` + `last_frame_url`), and reference-to-video
-(`reference_image_urls`). Images take one reference (`image_url`) or several
+Video takes five modes — text-to-video, image-to-video (`image_url`),
+first-and-last-frame (`image_url` + `last_frame_url`), reference-to-video
+(`reference_image_urls`), and Kling motion control (`image_url` + `video_url`:
+the character performs the clip's movement). Images take one reference (`image_url`) or several
 (`reference_image_urls`). Both take a `resolution`: `480p`…`4k` for video,
 `1k`/`2k`/`4k` for images, and a tier above a model's base rate costs more.
 
-The GPT Image family reads a `quality` on top of that — `low`, `medium` (the
+The GPT Image family and `ideogram-4.5` read a `quality` on top of that — `low`, `medium` (the
 default every catalog rate is measured at), `high`, and on `gpt-image-2.5` and
 `gpt-image-2.5-flare` also `xhigh` and `max`. It is how much compute the model
 spends, and the ladder is steep: `max` bills sixteen times `medium`.

package/cli/ottoport.mjs

@@ -253,8 +253,8 @@ async function download(url, out) {
  * what a shell user reaches for anyway, so both spellings are accepted:
  * `--ref a.png --ref b.png` and `--ref a.png,b.png`.
  */
-function references(flags) {
-  const raw = flags.ref ?? flags.reference;
+function references(flags, name = "ref") {
+  const raw = name === "ref" ? flags.ref ?? flags.reference : flags[name];
   const list = (Array.isArray(raw) ? raw : [raw])
     .filter(Boolean)
     .flatMap((value) => String(value).split(",").map((one) => one.trim()))
@@ -308,6 +308,8 @@ async function cmdVideo(prompt, flags) {
     ...(last(flags.image) ? { image_url: flags.image } : {}),
     ...(last(flags["last-frame"]) ? { last_frame_url: flags["last-frame"] } : {}),
     ...(references(flags) ? { reference_image_urls: references(flags) } : {}),
+    ...(references(flags, "ref-video") ? { reference_video_urls: references(flags, "ref-video") } : {}),
+    ...(references(flags, "ref-audio") ? { reference_audio_urls: references(flags, "ref-audio") } : {}),
     ...webhook(flags),
   };
   console.error("submitting video job (this can take a minute)…");
@@ -410,7 +412,7 @@ async function cmdAudio(kind, prompt, flags) {
 const HELP = `OttoPort — one command for every model.
 
 Usage:
-  ottoport models [--modality chat|image|video|tts|music] [--json]
+  ottoport models [--modality chat|image|video|tts|stt|music|embedding] [--json]
   ottoport models <model-id> [--json]     what that one model accepts
   ottoport chat  "<prompt>" [--model claude-sonnet-5] [--system "..."] [--no-stream]
                             [--video <url|file>[,<url|file>…]]   ask about a video
@@ -423,6 +425,7 @@ Usage:
   ottoport video "<prompt>" [--model kling-3.0] [--duration 5] [--resolution 1080p]
                             [--aspect-ratio 16:9] [--seed 7] [--image <url>]
                             [--last-frame <url>] [--ref <url>[,<url>…]]
+                            [--ref-video <url>[,<url>…]] [--ref-audio <url>[,<url>…]]
                             [--webhook-url https://…] [--webhook-secret <s>]
                             [--no-wait] [--out clip.mp4]
   ottoport video --model video-depth-anything --video <url> [--out depth.mp4]

package/mcp/server.mjs

@@ -218,7 +218,8 @@ async function generateImage({ prompt, model, ...rest }) {
 
 async function generateVideo({ prompt, model, webhook_url, webhook_secret, ...rest }) {
   // A depth model takes only `video_url`; the gateway says which models need a prompt.
-  if (!prompt && !rest.video_url) throw new Error("`prompt` is required");
+  // A draft completion takes everything from its draft, the prompt included.
+  if (!prompt && !rest.video_url && !rest.draft_id) throw new Error("`prompt` is required");
   // The secret is best left in the server's environment, so an agent never has
   // to carry it through a conversation.
   const secret = webhook_secret || process.env.OTTOPORT_WEBHOOK_SECRET;
@@ -241,6 +242,8 @@ async function generateVideo({ prompt, model, webhook_url, webhook_secret, ...re
   const { job, timedOut, path } = await waitForJob(VIDEO_JOB_PATH, submitted, { timeoutMs: 10 * 60_000, label: "video" });
   if (timedOut) return `Video job ${job.id} is still processing. Poll ${path} for its result.`;
   const url = job.data?.[0]?.url;
+  // The draft's job id is what completes it, so it is worth saying.
+  if (url && rest.draft) return `${url}\nDraft ${job.id} — complete it at 1080p with draft_id "${job.id}" within seven days.`;
   return url || JSON.stringify(job);
 }
 
@@ -284,7 +287,7 @@ const TOOLS = [
     inputSchema: {
       type: "object",
       properties: {
-        modality: { type: "string", enum: ["chat", "image", "video", "tts", "music"], description: "Filter to one modality." },
+        modality: { type: "string", enum: ["chat", "image", "video", "tts", "stt", "music", "embedding"], description: "Filter to one modality." },
         model: { type: "string", description: "A model id, e.g. veo-3.1. Returns that one model's full parameter list instead of the catalog." },
       },
     },
@@ -313,7 +316,7 @@ const TOOLS = [
       type: "object",
       properties: {
         prompt: { type: "string", description: "What to generate. On a layer-decomposition model, an optional caption of the input image." },
-        model: { type: "string", description: "Model id, e.g. gpt-image-2, nano-banana-2, nano-banana-pro. Default gpt-image-2." },
+        model: { type: "string", description: "Model id, e.g. gpt-image-2, nano-banana-2.1, nano-banana-pro. Default gpt-image-2." },
         size: { type: "string", description: 'Exact pixels, e.g. "1024x1024". Wins over `resolution`.' },
         // No `enum`: the vocabulary is "1k"/"2k"/"4k", but which of them a given
         // model actually offers differs per model, and an enum here reads as a
@@ -322,7 +325,7 @@ const TOOLS = [
         // at best. The model's own menu is in `ottoport_list_models`.
         resolution: { type: "string", description: 'Output detail — "1k", "2k" or "4k", but only the tiers this model lists in ottoport_list_models. Priced accordingly. Pair with `aspect_ratio` when you know the shape but not the pixel vocabulary; an explicit `size` wins.' },
         aspect_ratio: { type: "string", description: 'e.g. "16:9". Read only when `size` is absent.' },
-        quality: { type: "string", description: 'How much compute the model spends, on the GPT Image family only — "low", "medium" (the default), "high", and on gpt-image-2.5 "xhigh" and "max". Priced by tier, and the ladder is steep: `max` costs sixteen times `medium`. Check the model\'s `qualities` in ottoport_list_models.' },
+        quality: { type: "string", description: 'How much compute the model spends, on the GPT Image family and ideogram-4.5 — "low", "medium" (the default), "high", and on gpt-image-2.5 "xhigh" and "max". Priced by tier, and the ladder is steep: `max` costs sixteen times `medium`. Check the model\'s `qualities` in ottoport_list_models.' },
         n: { type: "number" },
         seed: { type: "number" },
         image_url: { type: "string", description: "One reference image, for editing and image-to-image — or, on a layer-decomposition model, the image to split." },
@@ -337,24 +340,30 @@ const TOOLS = [
   },
   {
     name: "ottoport_generate_video",
-    description: "Generate a video through an OttoPort video model (Veo, Kling, Seedance, …). Modes: text-to-video, image-to-video (`image_url`), first-and-last-frame (plus `last_frame_url`), reference-to-video (`reference_image_urls`), and on some models video edit/extend (`video_url`). video-depth-anything instead turns `video_url` into a grayscale depth video and takes no prompt. Which a model accepts is in its capabilities — call ottoport_list_models. Blocks on the queue and returns a video URL, unless `webhook_url` is given.",
+    description: "Generate a video through an OttoPort video model (Veo, Kling, Seedance, …). Modes: text-to-video, image-to-video (`image_url`), first-and-last-frame (plus `last_frame_url`), reference-to-video (`reference_image_urls`, plus `reference_video_urls` / `reference_audio_urls` on models whose capabilities list `referenceMedia`), and on some models video edit/extend (`video_url`). Seedance 2.5 also drafts: `draft: true` renders a cheap 480p preview, and `draft_id` set to that job's id renders the same shot at 1080p. video-depth-anything instead turns `video_url` into a grayscale depth video and takes no prompt. Kling motion control (`kling-2.6-motion-control`, `kling-3.0-motion-control` and their `-pro` variants) makes the character in `image_url` perform the movement in `video_url`; the prompt is optional and the output runs as long as the clip. Which a model accepts is in its capabilities — call ottoport_list_models. Blocks on the queue and returns a video URL, unless `webhook_url` is given.",
     inputSchema: {
       type: "object",
       properties: {
-        prompt: { type: "string", description: "Required, except on video-depth-anything, which takes none." },
-        model: { type: "string", description: "Model id, e.g. veo-3.1, kling-3.0, seedance-2.0-fast, video-depth-anything. Default kling-3.0." },
+        prompt: { type: "string", description: "Required, except on video-depth-anything, which takes none, and the motion-control models, where it is optional." },
+        model: { type: "string", description: "Model id, e.g. veo-3.1, kling-3.0, seedance-2.0-fast, kling-3.0-motion-control, video-depth-anything. Default kling-3.0." },
         duration: { type: "number", description: "Seconds. The main lever on what this costs." },
         // Per-model, not universal — see the note on the image tool. kling-3.0
         // has no 480p and no 4k; gemini-omni-flash has only 720p.
         resolution: { type: "string", description: 'Output detail — "480p", "720p", "1080p", "2k" or "4k", but only the tiers this model lists in ottoport_list_models. Priced accordingly, so it is worth reading before spending.' },
         aspect_ratio: { type: "string", description: 'e.g. "16:9".' },
         seed: { type: "number" },
-        image_url: { type: "string", description: "First frame, for image-to-video." },
+        image_url: { type: "string", description: "First frame, for image-to-video — or, on a motion-control model, the character who performs the movement." },
         last_frame_url: { type: "string", description: "Final frame. With `image_url`, this is first-and-last-frame generation." },
         reference_image_urls: { type: "array", items: { type: "string" }, description: "Subjects the shot carries through — a character, a product, a style — without being a frame of it." },
+        reference_video_urls: { type: "array", items: { type: "string" }, description: "Reference videos — motion, a performance, a scene — on models listing `referenceMedia` (Seedance 2.x, MiniMax H3, Wan 3.0). Billed per second of reference video where the upstream charges for it. To change one clip rather than draw on it, send it as `video_url` on a model listing video-edit." },
+        reference_audio_urls: { type: "array", items: { type: "string" }, description: "Reference audio — a voice, a beat — beside reference images or videos, on the same models. Not billed." },
         keyframes: { type: "array", items: { type: "object", properties: { image_url: { type: "string" }, time: { type: "number", description: "Seconds from the start; omit to spread evenly." } }, required: ["image_url"] }, description: "Images pinned in time, for keyframes-to-video." },
-        video_url: { type: "string", description: "A video to edit or extend, on models listing video-edit / video-extend — or the only input to video-depth-anything. Edits and depth bill per second of this video (at most 60s)." },
+        video_url: { type: "string", description: "A video to edit or extend, on models listing video-edit / video-extend — or the only input to video-depth-anything, or the movement a motion-control model transfers (3-30s; 3-10s with `character_orientation` \"image\"). Edits, depth and motion control bill per second of this video (edits and depth at most 60s; motion control at least 3s)." },
+        character_orientation: { type: "string", enum: ["video", "image"], description: "Motion-control models only: whether the character faces the way the person in `video_url` does (\"video\", the default) or keeps the pose of `image_url` (\"image\")." },
+        keep_original_sound: { type: "boolean", description: "Motion-control models only: keep the audio of `video_url` in the result." },
         video_task: { type: "string", enum: ["edit", "extend"], description: "edit (default) restyles video_url; extend continues it for `duration` more seconds." },
+        draft: { type: "boolean", description: "Seedance 2.5 only: render a 480p preview at about a quarter of the 1080p price, to check the shot before paying for it. Text, image and reference-image requests; no `resolution`." },
+        draft_id: { type: "string", description: "Seedance 2.5 only: the job id of a finished draft, to render that same shot at 1080p. Send `model` and this alone — prompt, inputs and duration come from the draft. Within seven days of the draft." },
         webhook_url: { type: "string", description: "Public HTTPS endpoint to POST the finished job to (events video.completed / video.failed, signed with HMAC-SHA256). With it the tool returns the job id at once instead of waiting." },
         webhook_secret: { type: "string", description: "Signing secret for webhook_url. Omit to use the server's OTTOPORT_WEBHOOK_SECRET." },
       },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ottoport",
-  "version": "1.9.0",
+  "version": "1.11.0",
   "description": "Claude Code plugin, CLI and MCP server for OttoPort — one OpenAI-compatible API for every LLM, image, video, and speech model.",
   "homepage": "https://ottoport.ai",
   "repository": {

package/skills/ottoport/SKILL.md

@@ -16,8 +16,8 @@ and do not call the HTTP API, when a tool covers the job:
 | --- | --- |
 | `ottoport_list_models` | The catalog: what each model is for, and its rate in credits and USD. Filter with `modality`, or pass `model` for ONE model in full — its modes, its resolution tiers with prices, and every parameter it reads. |
 | `ottoport_chat` | `prompt`, plus optional `model`, `system`, `temperature`, `max_tokens`. |
-| `ottoport_generate_image` | `prompt`, plus optional `model`, `size` or `resolution`, `quality` (GPT Image only), `aspect_ratio`, `n`, `seed`, `image_url` (one reference), `reference_image_urls` (several). Layer decomposition too: `model` `qwen-image-layered` or `seedream-5.0-layers` with `image_url` alone (prompt optional, `num_layers` on Qwen) returns one URL per RGBA layer. |
-| `ottoport_generate_video` | `prompt`, plus optional `model`, `duration`, `resolution`, `aspect_ratio`, `seed`, `image_url` (first frame), `last_frame_url`, `reference_image_urls`, `webhook_url` (returns the job id at once and POSTs the result there; the secret comes from `webhook_secret` or the server's `OTTOPORT_WEBHOOK_SECRET`). |
+| `ottoport_generate_image` | `prompt`, plus optional `model`, `size` or `resolution`, `quality` (GPT Image and Ideogram), `aspect_ratio`, `n`, `seed`, `image_url` (one reference), `reference_image_urls` (several). Layer decomposition too: `model` `qwen-image-layered` or `seedream-5.0-layers` with `image_url` alone (prompt optional, `num_layers` on Qwen) returns one URL per RGBA layer. |
+| `ottoport_generate_video` | `prompt`, plus optional `model`, `duration`, `resolution`, `aspect_ratio`, `seed`, `image_url` (first frame), `last_frame_url`, `reference_image_urls`, `reference_video_urls` / `reference_audio_urls` (models listing `referenceMedia`), `video_url` (edit / extend, or the movement for motion control), `character_orientation` / `keep_original_sound` (motion control), `draft` / `draft_id` (Seedance 2.5: a 480p preview, then the same shot at 1080p from the preview's job id), `webhook_url` (returns the job id at once and POSTs the result there; the secret comes from `webhook_secret` or the server's `OTTOPORT_WEBHOOK_SECRET`). |
 | `ottoport_video_webhook` | `job_id`, plus `action` `status` (every delivery attempt and the endpoint's response) or `retry` (redeliver a finished job's webhook now). |
 | `ottoport_generate_speech` | `prompt` (the text), plus optional `model`, `voice`, `format`. |
 | `ottoport_generate_music` | `prompt`, plus optional `model`, `duration`, `format`. |
@@ -28,7 +28,7 @@ say so rather than reaching for a shell. `ottoport install <host>` registers it
 
 ## Modes
 
-Video takes four, and which ones a model accepts is per-model — `ottoport_list_models`
+Video takes five, and which ones a model accepts is per-model — `ottoport_list_models`
 prints each one's menu beside its price:
 
 | Mode | Send |
@@ -36,7 +36,8 @@ prints each one's menu beside its price:
 | text-to-video | `prompt` alone |
 | image-to-video | `image_url` — the first frame |
 | first-and-last-frame | `image_url` **and** `last_frame_url` |
-| reference-to-video | `reference_image_urls` — a character, a product, a style the shot carries through without either being a frame of it |
+| reference-to-video | `reference_image_urls` — a character, a product, a style the shot carries through without either being a frame of it; plus `reference_video_urls` and `reference_audio_urls` (motion, a performance, a voice) on models whose capabilities list `referenceMedia` — Seedance 2.x, MiniMax H3, Wan 3.0 |
+| motion-control | `image_url` (the character) **and** `video_url` (a 3-30s clip of the movement), on `kling-2.6-motion-control`, `kling-3.0-motion-control` and their `-pro` variants. Prompt optional; `character_orientation` `"video"` (default) or `"image"` (clip then capped at 10s), `keep_original_sound`. Bills per second of the clip, 3s at least |
 
 Images are the same idea with one axis: `image_url` for a single edit, or
 `reference_image_urls` for several at once, up to that model's limit.
@@ -55,8 +56,8 @@ video and `"1k"`/`"2k"`/`"4k"` for images, but **no model offers all of it**:
 `gpt-image-1.5` only 1k. Same for how many references a model takes —
 `nano-banana-lite` accepts 3, `gpt-image-2` accepts 16.
 
-`quality` is a second priced dial, on the GPT Image family alone: `low`,
-`medium` (the default, and the tier every catalog rate was measured at),
+`quality` is a second priced dial, on the GPT Image family and `ideogram-4.5`:
+`low`, `medium` (the default, and the tier every catalog rate was measured at),
 `high`, and `xhigh`/`max` on `gpt-image-2.5` and `gpt-image-2.5-flare`. It buys
 compute, not pixels, and the ladder is steep — `max` bills sixteen times
 `medium`. A model's tiers are its `qualities` in `ottoport_list_models`.
@@ -77,7 +78,7 @@ sensible default.
 | Modality | Default | Also available |
 | --- | --- | --- |
 | chat | `claude-sonnet-5` | `claude-haiku-4.5` (cheap), `gpt-5.5`, `gemini-3.5-flash` |
-| image | `gpt-image-2` | `nano-banana-pro` (high fidelity), `nano-banana-2` (edits) |
+| image | `gpt-image-2` | `nano-banana-2.1` (fast generation and edits), `nano-banana-pro` (high fidelity) |
 | video | `kling-3.0` | `seedance-2.0-fast` (cheap), `veo-3.1` (with audio) |
 | speech | `gpt-4o-mini-tts` | `eleven-v3` |
 | music | `suno-v5` | `lyria-2` |