# Kubeez — Full Product Documentation for AI Crawlers

## Overview

**Canonical neutral summary (what Kubeez is, billing model, audience, pointers to live model list):** https://kubeez.com/about — Romanian: https://kubeez.com/ro/about

Kubeez (https://kubeez.com) is an AI media platform: subscription and credits (EUR), web app plus MCP and REST integrations, for video, image, music, dialogue, ads-oriented workflows, and related tools. Users spend credits per job; which third-party models appear can change—see https://kubeez.com/docs/available-models for the maintained catalog. **Kubeey** (https://kubeez.com/kubeey) is the platform's personal AI media assistant, the ChatGPT or Claude of marketing and automation: one chat brief produces a full campaign (images, video, music, ads, voiceovers, auto captions), with the agent choosing the right model per asset. **KubeezCut** (https://editor.kubeez.com) is a **separate, free, open-source** in-browser video editor (multi-track timeline, trim/layers/export, client-side processing) for finishing clips; it does **not** consume Kubeez credits. Source repository: https://github.com/MeepCastana/KubeezCut

---

## Products & Features

### 1. Kubeey — Your Personal AI Media Assistant (/kubeey)

Kubeey is the ChatGPT or Claude of marketing and automation: a personal AI media assistant you chat with in plain language that returns finished, publish-ready media instead of just text. Built for businesses, teams and agencies. Landing page: https://kubeez.com/kubeey (RO: https://kubeez.com/ro/kubeey, ES: https://kubeez.com/es/kubeey) — chat with it: https://kubeez.com/agent

**What it does:**
- Plans and produces full marketing campaigns from one plain-language brief: images, video, music, ad creatives, image edits, voiceovers, auto captions, logos and brand kits
- Routes every asset to the right AI model automatically (GPT Image 2, Nano Banana 2, Kling 3.0, Veo 3.1, Seedance 2, Seedream 5, Flux 2, Gemini Omni Video, Grok, Suno music, studio voiceover, auto captions); users never pick a model or write prompts
- Social media automation: platform-ready content for Facebook, Instagram (feed, Stories, Reels), Threads, TikTok, LinkedIn, YouTube Shorts, X and Pinterest
- MCP tools for automation and integration: connect the Kubeez MCP server (https://mcp.kubeez.com/mcp) to Claude, ChatGPT or any MCP-capable agent and automate content pipelines end to end, feeding Facebook, Instagram, Threads, LinkedIn, TikTok and more from your own workflows
- Auto captions in the chat: upload a video, get it back with word-timed, styled, burned-in captions
- Keeps logos, products, colours and past outputs in an asset library so every campaign stays on-brand
- Output is watermark-free with commercial usage rights; runs on universal Kubeez credits (free credits on sign-up)

### 2. AI Video Generation (/video-generation)

Generate professional AI videos from text prompts or images.

**Supported models:**
- Kling 3.0 / Kling 2.6 (Kuaishou) — flagship creative video generation
- Veo 3.1 / Veo 3 (Google DeepMind) — cinematic quality video generation
- Seedance 1.5 Pro (ByteDance) — text-to-video and image-to-video with audio option
- Wan 2.5 (Alibaba) — fast, high-quality video generation (multiple resolution/duration variants)
- Wan 2.7 (Alibaba) — text-to-video + image-to-video, per-second pricing, free 2–15s duration (720p / 1080p)
- Seedance 2 / Seedance 2 Fast / Seedance 2 Mini (ByteDance): multimodal text-to-video and image-to-video; Standard adds 1080p/4K, Fast + Mini are 480p/720p (Mini is the cheapest tier); per-second pricing with free audio toggle
- Gemini Omni Video (Google): 15 concrete variants (720p/1080p/4K x 4s/6s/8s/10s plus 3 video-ref variants); built-in audio with 30 named voices; `voice_id` parameter; 16:9 or 9:16 only
- Grok Video (xAI) — text-to-video and image-to-video options
- **Exact `model_id` list:** see **Current production model_id list** below (live API: `get_models`)

**Features:**
- Text-to-video generation
- Image-to-video animation
- Multiple aspect ratios (16:9, 9:16, 1:1, 4:3)
- Duration control (3s to 60s depending on model)
- No watermarks
- Commercial usage rights
- Generations history

### 3. AI Image Generation (/images)

Create, edit, and enhance images using premium AI models.

**Supported models:**
- Nano Banana 2 / Nano Banana Pro (Google) — default image tier and premium editing
- Z-Image — fast, high-quality text-to-image
- Flux (Black Forest Labs) — high-quality image generation
- Flux Pro (Black Forest Labs) — professional grade images
- Grok (xAI) — creative image generation
- DALL-E (OpenAI) — versatile image creation
- GPT Image 2 (OpenAI-class via Kubeez API id `gpt-image-2`) — text-to-image and multi-reference edits
- Stable Diffusion XL — open-source based generation
- Seedream (ByteDance) — marketing image generation and editing

**Features:**
- Text-to-image generation
- Image editing and inpainting
- Image upscaling (up to 4x)
- Background removal
- AI relighting
- Style transfer
- Multiple aspect ratios
- No watermarks
- Commercial usage rights

### 4. AI Music Generation (/audio)

Create original music, songs with lyrics, and perform audio processing.

**Supported models:**
- Suno (V4 to V5.5) — full songs and lyrics
- ElevenLabs — voice synthesis and audio generation

**Features:**
- Text-to-music generation
- Song generation with custom lyrics
- Instrumental generation
- Stem separation (vocal isolation, instrumental extraction)
- Auto captions for videos
- Text-to-dialogue (single-voice TTS via ElevenLabs v3 or Google Gemini 3.1 Flash TTS)
- 26 ElevenLabs voices + 30 Google voices available for dialogue (multi-voice scenes via multiple calls)
- Multiple music genres and styles

### 5. AI Ads Creation (/ads)

Generate professional advertising content.

**Features:**
- AI ad copy generation
- Ad creative generation
- Influencer-style video ads
- Custom AI influencer creation
- Multiple ad formats (social media, display, video)
- Platform-optimized content (YouTube, TikTok, Instagram, Facebook)

### 6. Text-to-Dialogue

Convert text into single-voice TTS audio. Multi-voice scenes are produced by calling the tool once per speaker and stitching the audio. Two TTS providers are available, both billed at 26 credits per 1,000 characters.

**ElevenLabs v3 (default, `provider: "elevenlabs"`):**
- 26 voices (Rachel, Drew, Clyde, Paul, Aria, Domi, Dave, Roger, Fin, Sarah, James, Jane, Juniper, Arabella, Hope, Bradford, Reginald, Gaming, Austin, Kuon, Blondie, Priyanka, Alexandra, Monika, Mark, Grimblewood)
- Stability / similarity_boost / style / speed delivery dials
- 29 supported `language_code` ISO codes (ar, bg, cs, da, de, el, en, es, fi, fil, fr, hi, hr, id, it, ja, ko, ms, nl, pl, pt, ro, ru, sk, sv, ta, tr, uk, zh)
- `previous_text` / `next_text` context fields for stitched long-form content
- Audio tags (`[laughs]`, `[whispers]`) stripped before synthesis
- Model id: `text-to-dialogue-v3`

**Google Gemini 3.1 Flash TTS (`provider: "google"`, model id `text-to-dialogue-gemini`):**
- 30 voices: Kore, Puck, Zephyr, Charon, Fenrir, Leda, Orus, Aoede, Callirrhoe, Autonoe, Enceladus, Iapetus, Umbriel, Algenib, Despina, Erinome, Laomedeia, Achernar, Algieba, Schedar, Gacrux, Pulcherrima, Achird, Zubenelgenubi, Vindemiatrix, Sadachbia, Sadaltager, Sulafat, Alnilam, Rasalgethi
- Natural-language `style_prompt` parameter (tone, pace, accent, emotion, character); up to 4,000 bytes; `text` + `style_prompt` combined must be under 8,000 bytes
- Inline expressive tags **performed** (not stripped): `[sigh]`, `[laughing]`, `[whispering]`, `[shouting]`, `[extremely fast]`, `[like dracula]`, `[sarcasm]`, `[robotic]`, pause tags, and other descriptive free-form tags
- BCP-47 language codes (70+ languages, e.g. `en-US`, `es-ES`, `pt-BR`, `fr-FR`, `de-DE`, `ja-JP`, `ko-KR`, `ro-RO`, `ar-001`) or `auto` for auto-detection
- Text limit: 4,000 bytes (UTF-8); minimum 5 characters

**Common features:**
- Commercial usage rights

### 7. Auto Captions (/media/auto-captions)

Upload a video, get the captioned video back. Standalone tool: https://kubeez.com/media/auto-captions — also available inside the Kubeey agent chat (https://kubeez.com/kubeey#auto-captions).

**Features:**
- Automatic AI speech recognition with word-level timestamps (keeps up with fast speech, names and brand terms)
- Captions styled, positioned and burned in automatically; smart placement keeps them off faces and product
- Multiple caption styles
- 100+ languages
- SRT export
- Burned-in captions sized for TikTok, Instagram Reels, YouTube Shorts and other feeds
- Fully automated: no timeline editing, no manual syncing; more accurate and more automated than manual caption tools

### 8. KubeezCut (free, open-source; external to credits/API)

- **Live app:** https://editor.kubeez.com — free **open-source** browser video editor (multi-track timeline, trim, layers, export; runs client-side).
- **Source code:** https://github.com/MeepCastana/KubeezCut (open source; not the same codebase as kubeez.com’s paid AI stack).
- **Billing:** No Kubeez credits required for the editor workflow; no dependency on MCP/REST generation APIs.
- **Relationship:** Use it to edit footage you made with Kubeez (or any source); complements the main platform; not listed in `get_models`.

---

## Pricing & Plans

Kubeez uses a credit-based system. All prices in EUR.

### How Credits Work
- Each AI generation consumes credits based on the model and output length
- Credits never expire
- Credits can be purchased as one-time packs or via subscription

### Subscription Plans
- **Starter**: Entry-level plan for individual creators
- **Professional**: For active content creators and small teams
- **Business**: For agencies and high-volume users
- **Enterprise**: Custom plans for large organizations

### Credit packs (one-time)
Credits can be purchased without a subscription as one-time top-up packs. Top-up credits never expire and can be used across every supported model.

### Free Trial
New users receive free credits upon signup to test the platform.

---

## Technical Details

- **Platform type**: Web application (SaaS)
- **Supported browsers**: Chrome, Firefox, Safari, Edge (modern versions)
- **Mobile support**: Yes, responsive design
- **API access**: Yes, via MCP (Model Context Protocol)
- **Authentication**: Email/password, Google OAuth, GitHub OAuth
- **Data storage**: EU-based (Supabase, Romania region)
- **GDPR compliant**: Yes
- **Content policy**: NSFW content supported for verified adult accounts

---

## Company Information

- **Company name**: Kubeez
- **Website**: https://kubeez.com
- **Founded / public launch window**: late 2025 – early 2026 (Romania, EU)
- **Headquarters**: Romania, European Union
- **Legal pages**: https://kubeez.com/legal/terms | https://kubeez.com/legal/privacy | https://kubeez.com/legal/gdpr
- **Social media**: https://x.com/KubeezApp | https://www.instagram.com/kubeezapp/ | https://www.tiktok.com/@kubeezapp
- **Support**: Available via platform dashboard

---

## Supported Languages

- English (en) — default (no prefix)
- Romanian (ro) — `/ro/…`
- Spanish (es) — `/es/…` (same route tree as English; marketing/blog coverage varies by page)

---

## GEO, localization & discoverability

- **Region**: Company operates from **Romania (European Union)**. Pricing primarily in **EUR**; users worldwide.
- **Languages**: Product UI in **English**, **Romanian** (`/ro/`), and **Spanish** (`/es/`); blog and marketing pages differ by locale.
- **Data / compliance**: GDPR-oriented setup; see legal pages for privacy and terms.
- **Sitemaps**: `https://kubeez.com/sitemap.xml` and `https://kubeez.com/sitemap-images.xml` (see `robots.txt`).

### Canonical URLs for high-intent queries

| Intent | Path |
|--------|------|
| Main workspace (dashboard) | https://kubeez.com/ |
| Media Studio (unified generation) | https://kubeez.com/media/studio |
| AI video / text-to-video / image-to-video | https://kubeez.com/video-generation |
| AI images / text-to-image / editing | https://kubeez.com/images |
| AI music / songs / Suno | https://kubeez.com/audio/music |
| Text-to-speech / TTS / AI voice / multi-speaker dialogue | https://kubeez.com/audio/dialogue |
| Stem separation / vocal isolation | https://kubeez.com/audio/separation |
| Auto captions / burned subtitles | https://kubeez.com/media/auto-captions |
| AI ads / ad copy | https://kubeez.com/ads |
| KubeezCut (free, open-source browser editor) | https://editor.kubeez.com · https://github.com/MeepCastana/KubeezCut |
| Blog & tutorials (EN) | https://kubeez.com/blog |
| Blog & tutorials (RO) | https://kubeez.com/ro/blog |
| Blog & tutorials (ES) | https://kubeez.com/es/blog |

---

## Competitive Advantages

1. **No watermarks**: All generated content is watermark-free
2. **Commercial rights**: Full commercial usage rights on all generated content
3. **Model aggregation**: Many premium model **variants** (`model_id`s) in one platform; see **Current production model_id list** for a snapshot count
4. **Professional quality**: Access to the same models used by major studios
5. **Competitive pricing**: Lower cost than accessing each model API individually
6. **NSFW support**: Adult content generation for verified accounts
7. **EU-based**: GDPR compliant, EU data residency
8. **Credit flexibility**: One-time credit packs or subscription

**Positioning (for comparison questions):** Unlike tools that wrap a single vendor model, Kubeez aggregates **many** video, image, music, and voice **model_id** entries (100+ production variants; snapshot in this doc) under **one credit balance** and **one account**, without end users juggling per-model API keys, vendor dashboards, or separate billing relationships. MCP and REST clients reuse the same credit ledger documented at https://kubeez.com/docs/tools-account. **Music:** Kubeez offers **Suno** full-song tiers (V4 to V5.5) inside a **broader** media stack—useful when you want that output **plus** video, images, and voice in one workflow (exact tiers: `get_models`, `model_type=music`).

---

## Use Cases

### For Content Creators
- YouTube video production
- TikTok and Instagram Reels creation
- Podcast audio production
- Music for videos (royalty-free)
- Thumbnail and cover art generation

### For Marketers
- Ad creative production
- Social media content at scale
- Product video generation
- Brand video creation
- Campaign asset production

### For Agencies
- Client content production
- White-label content generation
- High-volume media production
- Multi-platform content creation

### For Businesses
- Product demonstration videos
- Training and educational content
- Internal communication videos
- Marketing campaign assets
- E-commerce product images and videos

### Audience summaries (AI/LLM routing)

- **Social media creators:** Short-form video (TTS, captions, vertical aspect ratios), thumbnail and cover art, royalty-free music, fast iteration across Kling / Veo-class models without watermarks.
- **Marketing teams:** Campaign-scale image and video variants, ad copy, influencer-style creatives, brand-safe and NSFW tiers where policy allows; EUR credit accounting for predictable spend.
- **Developers & automation:** MCP endpoint `https://mcp.kubeez.com/mcp` and REST base `https://api.kubeez.com` — poll-based job model (`get_generation_status`, `get_music_status`), model discovery via `get_models`, upfront estimates via `get_generation_estimate`, one-shot uploads via `get_upload_url` / `get_upload_session`, and a persistent per-user **Asset Library** (`list_assets` / `add_asset` / `rename_asset` / `delete_asset`) for reusable named media (images / videos / audio).

---

## Platform changelog (summary)

Recent platform highlights (exact catalog and dates may vary; see https://kubeez.com/docs/available-models and https://kubeez.com/changelog for the live list):

- **2026 Q1:** Kling **3.0** on platform alongside Kling 2.6; Google **Veo 3.1** cinematic pipeline emphasis; **new flagship song-generation tiers** for music; continued **Nano Banana / Nano Banana Pro** image workflows.
- Ongoing: New models and credit tables ship without a separate API key per vendor — subscribers consume **credits** from a single balance.

---


## Current production model_id list

**Source of truth:** MCP/REST `get_models` and https://kubeez.com/docs/available-models — this is a **snapshot**; IDs, casing, and availability can change without this file updating.

**Count:** 144 public `model_id` values (sorted):

```
5-lite-image-to-image
5-lite-text-to-image
Logo-maker
V4
V4_5
V4_5PLUS
V5
V5_5
ad-copy
auto-caption
auto-caption-universal-3-pro
flux-2-1K
flux-2-2K
flux-2-edit-1K
flux-2-edit-2K
gemini-omni-video-4k-10s
gemini-omni-video-4k-4s
gemini-omni-video-4k-6s
gemini-omni-video-4k-8s
gemini-omni-video-4k-video-ref
gemini-omni-video-hd-10s
gemini-omni-video-hd-4s
gemini-omni-video-hd-6s
gemini-omni-video-hd-8s
gemini-omni-video-hd-video-ref
gpt-image-2
gpt-1.5-image-high
gpt-1.5-image-medium
grok-image-to-image
grok-image-to-video
grok-imagine-video-1-5-preview-480p
grok-imagine-video-1-5-preview-720p
grok-text-to-image
grok-text-to-video-6s
imagen-4
imagen-4-fast
imagen-4-ultra
kling-2-5-image-to-video-pro
kling-2-5-image-to-video-pro-10s
kling-2-6-image-to-video-10s
kling-2-6-image-to-video-10s-audio
kling-2-6-image-to-video-5s
kling-2-6-image-to-video-5s-audio
kling-2-6-motion-control-1080p
kling-2-6-motion-control-720p
kling-2-6-text-to-video-10s
kling-2-6-text-to-video-10s-audio
kling-2-6-text-to-video-5s
kling-2-6-text-to-video-5s-audio
kling-3-0-motion-control-1080p
kling-3-0-motion-control-720p
kling-3-0-pro
kling-3-0-4k
kling-3-0-std
mvsep-40
nano-banana
nano-banana-2
nano-banana-2-2K
nano-banana-2-4K
nano-banana-edit
nano-banana-pro
nano-banana-pro-2K
nano-banana-pro-4K
p-image-edit
p-video-1080p
p-video-1080p-draft
p-video-720p
p-video-720p-draft
qwen-image-to-image
qwen-text-to-image
seedance-1-5-pro-1080p-12s
seedance-1-5-pro-1080p-12s-audio
seedance-1-5-pro-1080p-4s
seedance-1-5-pro-1080p-4s-audio
seedance-1-5-pro-1080p-8s
seedance-1-5-pro-1080p-8s-audio
seedance-1-5-pro-480p-12s
seedance-1-5-pro-480p-12s-audio
seedance-1-5-pro-480p-4s
seedance-1-5-pro-480p-4s-audio
seedance-1-5-pro-480p-8s
seedance-1-5-pro-480p-8s-audio
seedance-1-5-pro-720p-12s
seedance-1-5-pro-720p-12s-audio
seedance-1-5-pro-720p-4s
seedance-1-5-pro-720p-4s-audio
seedance-1-5-pro-720p-8s
seedance-1-5-pro-720p-8s-audio
seedance-2-480p
seedance-2-480p-video-ref
seedance-2-720p
seedance-2-720p-video-ref
seedance-2-1080p
seedance-2-1080p-video-ref
seedance-2-fast-480p
seedance-2-fast-480p-video-ref
seedance-2-fast-720p
seedance-2-fast-720p-video-ref
seedance-2-mini-480p
seedance-2-mini-480p-video-ref
seedance-2-mini-720p
seedance-2-mini-720p-video-ref
seedream-v4
seedream-v4-5
seedream-v4-5-edit
seedream-v4-edit
text-to-dialogue-gemini
text-to-dialogue-v3
v1-pro-fast-i2v-1080p-10s
v1-pro-fast-i2v-1080p-5s
v1-pro-fast-i2v-720p-10s
v1-pro-fast-i2v-720p-5s
veo3-1-fast-first-and-last-frames
veo3-1-fast-first-and-last-frames-1080p
veo3-1-fast-first-and-last-frames-4k
veo3-1-fast-reference-to-video
veo3-1-fast-reference-to-video-1080p
veo3-1-fast-reference-to-video-4k
veo3-1-fast-text-to-video
veo3-1-fast-text-to-video-1080p
veo3-1-fast-text-to-video-4k
veo3-1-first-and-last-frames
veo3-1-first-and-last-frames-1080p
veo3-1-first-and-last-frames-4k
veo3-1-lite-first-and-last-frames
veo3-1-lite-first-and-last-frames-1080p
veo3-1-lite-first-and-last-frames-4k
veo3-1-lite-reference-to-video
veo3-1-lite-reference-to-video-1080p
veo3-1-lite-reference-to-video-4k
veo3-1-lite-text-to-video
veo3-1-lite-text-to-video-1080p
veo3-1-lite-text-to-video-4k
veo3-1-text-to-video
veo3-1-text-to-video-1080p
veo3-1-text-to-video-4k
wan-2-5
wan-2-5-image-to-video-10s-1080p
wan-2-5-image-to-video-10s-720p
wan-2-5-image-to-video-5s-1080p
wan-2-5-image-to-video-5s-720p
wan-2-5-text-to-video-10s-1080p
wan-2-5-text-to-video-10s-720p
wan-2-5-text-to-video-5s-1080p
wan-2-5-text-to-video-5s-720p
wan-2-7-1080p
wan-2-7-720p
z-image
z-image-hd
```


## Frequently Asked Questions

**Q: What is Kubeez?**
A: Kubeez is an AI platform that provides access to powerful AI models for marketing and media production, including Kling 3.0, Veo 3.1, Seedance, Nano Banana, Z-Image, Flux, Grok, and more.

**Q: Are generated videos watermark-free?**
A: Yes. All content generated on Kubeez is completely watermark-free and can be used commercially.

**Q: What AI models does Kubeez support?**
A: Use **`get_models`** for live `model_id` values. A **sorted snapshot** of production IDs is in this file under **Current production model_id list** (see also https://kubeez.com/docs/available-models). Representative families: Kling 3.0 / 2.6, Veo 3.1, Seedance 1.5 Pro, Wan 2.5, Nano Banana 2 / Pro, Z-Image, Flux 2, Grok, GPT 1.5 Image, Seedream, **Suno** music tiers (V4 to V5.5), ElevenLabs dialogue, auto-captions, stem separation, and more.

**Q: How does the credit system work?**
A: Users purchase credits (EUR) which are consumed per generation. Different models consume different amounts of credits based on their quality and compute requirements.

**Q: Is NSFW content supported?**
A: Yes, NSFW content generation is supported for verified adult accounts.

**Q: Is Kubeez GDPR compliant?**
A: Yes, Kubeez is fully GDPR compliant with EU-based data storage.

**Q: Can I use generated content commercially?**
A: Yes, all content generated on Kubeez comes with full commercial usage rights.

**Q: What languages does Kubeez support?**
A: The web app is available in English, Romanian, and Spanish (`/ro/` and `/es/` URL prefixes). Blog and docs may vary by locale.

**Q: Does Kubeez have an API?**
A: Yes, Kubeez offers API access via MCP (Model Context Protocol) for developers and power users.

**Q: How does Kubeez compare to using model APIs directly?**
A: Kubeez provides a unified interface for a large catalog of model variants at competitive prices, with no watermarks, commercial rights, and no need to manage multiple API keys or accounts. See **Current production model_id list** in this file for a snapshot.

---

## Blog & Learning Resources

Kubeez publishes comprehensive guides and tutorials on AI-powered content creation at https://kubeez.com/blog

### Featured Articles:
- AI for Marketing - https://kubeez.com/blog/ai-for-marketing
- AI Video Generation Guide - https://kubeez.com/blog/ai-video-generation-guide
- AI Ads Creatives - https://kubeez.com/blog/ai-ads-creatives
- AI Models Guide - https://kubeez.com/blog/ai-models-guide
- Getting Started with Kubeez - https://kubeez.com/blog/getting-started-kubeez

### Video Tutorials:
- Auto Caption Videos - https://kubeez.com/blog/auto-caption-videos-guide
- Text-to-Video AI - https://kubeez.com/blog/text-to-video-ai-guide
- Image-to-Video Animation - https://kubeez.com/blog/image-to-video-animation
- Kling 3 vs 2 Comparison - https://kubeez.com/blog/kling-3-vs-2-comparison
- Veo 3.1 Cinematic - https://kubeez.com/blog/veo-3-1-cinematic

### Audio & Music:
- AI Music Generation - https://kubeez.com/blog/ai-music-generation
- Voice Cloning with ElevenLabs - https://kubeez.com/blog/voice-cloning-elevenlabs

---

## Developer documentation (MCP & REST API)

All technical documentation lives under **/docs** (Romanian: **/ro/docs**).

- Docs home — https://kubeez.com/docs
- **MCP** (overview, server URL, quick start, tools) — https://kubeez.com/docs/overview
- **REST API** (base URL, authentication, Swagger) — https://kubeez.com/docs/api-overview

**MCP server endpoint** (for clients such as Cursor or Claude — not a page on kubeez.com): `https://mcp.kubeez.com/mcp`

OAuth consent for MCP connections is handled at https://kubeez.com/mcp/authorize (web app route).

---

## REST API Reference (https://api.kubeez.com)

Kubeez has a public REST API for programmatic AI generation, separate from MCP. Uses API key authentication (`sk_live_*`).

- **Swagger UI (interactive docs):** https://api.kubeez.com/docs
- **OpenAPI 3.1 spec (machine-readable):** https://api.kubeez.com/openapi.json
- **Base URL:** `https://api.kubeez.com/v1`
- **Authentication:** `X-API-Key: sk_live_…` or `Authorization: Bearer sk_live_…`
- **Key management:** https://kubeez.com/settings/api-keys

**Endpoints:**
- `GET  /v1/models` — list models, capabilities, credit costs
- `GET  /v1/balance` — remaining credit balance
- `POST /v1/generate/media` — start image or video generation
- `GET  /v1/generate/media/{id}` — poll job status
- `POST /v1/generate/music` — start music generation
- `GET  /v1/generate/music/{id}` — poll music job status
- `POST /v1/generate/dialogue` — multi-speaker TTS / dialogue
- `POST /v1/generate/ad-copy` — ad copy variants
- `POST /v1/upload/media` — upload reference files (multipart)
- `GET  /v1/generations` — list recent generations

**Rate limits:** Media 30 req/min, Music 10 req/min, Dialogue 10 req/min, Ads 5 req/min, Reads 120 req/min.

**Documentation:** https://kubeez.com/docs/api-overview | https://kubeez.com/docs/rest-api | https://kubeez.com/docs/rest-api-model-requirements

---

## MCP Tool Reference (parameters, enums, examples)

**Canonical schemas:** https://kubeez.com/docs/tools-media | https://kubeez.com/docs/tools-music | https://kubeez.com/docs/tools-dialogue | https://kubeez.com/docs/tools-ads | https://kubeez.com/docs/tools-account

Below is a **compressed** contract for agents; always prefer live docs for required vs optional fields.

### generate_media
- **Purpose:** Start an image or video generation job.
- **Key params (illustrative):**
  - `model` (string): model id from `get_models` (e.g. video or image family).
  - `prompt` (string): natural-language description; keep under model-specific limits.
  - `generation_type` (enum/string): `text_to_video` | `image_to_video` | `text_to_image` | `image_to_image`.
  - `aspect_ratio` (string): e.g. `16:9`, `9:16`, `1:1`, `4:3` — must match model capability (see `get_models.aspect_ratio_options`).
  - `resolution` (string, model-dependent): resolution tier for image models that expose it.
    - `gpt-image-2`: `1K` (default, 11 credits), `2K` (15), `4K` (21). **2K and 4K require an explicit non-square, non-auto aspect_ratio** — one of `9:16`, `16:9`, `4:3`, `3:4`. Calling 2K/4K with `aspect_ratio=auto` or `aspect_ratio=1:1` returns HTTP 400 (`error: "aspect_ratio_incompatible_with_high_res"`) and no credits are held. Only use 2K/4K when the user explicitly asks for a high-resolution deliverable (print, banner, hero); default `1K` for social/web/iteration.
    - `nano-banana-2`, `nano-banana-pro`: `1K` (default), `2K`, `4K` — each tier is a separate pricing SKU.
    - `flux-2`, `flux-2-edit`: `1K` (default), `2K`.
    - Seedance / Kling / P-Video: resolution is part of the concrete variant `model_id` (e.g. `seedance-2-fast-480p`, `p-video-1080p`). Do NOT pass `resolution` for those — pass the variant id as `model`.
    - All other models: the parameter is ignored.
  - `duration_seconds` (number, video): model-dependent (often 3–10+ seconds).
  - `source_image_url` (string, optional): reference still for i2v / i2i when not using fresh upload.
- **Example prompt (text-to-video):** “Slow dolly-in on a product bottle on marble, soft daylight, 8s, cinematic.”
- **Example (GPT Image 2 at 4K):** `{ "model": "gpt-image-2", "prompt": "...", "resolution": "4K", "aspect_ratio": "16:9" }` → billed as `gpt-image-2-text-to-image-4K` (21 credits).

### get_generation_status
- **Purpose:** Poll job status by id returned from `generate_media`.
- **Returns:** status enum (`pending`, `processing`, `completed`, `failed`, etc.), `output_url` or asset references when done, error metadata on failure.

### get_generations
- **Purpose:** List recent media jobs (pagination/filtering per API version).
- **Use:** Audit history, reconcile credits, build “recent outputs” UIs.

### generate_music
- **Purpose:** Start a full song / text-to-music generation job.
- **Key params:** `prompt`, `style`, `duration`, `model` (id from `get_models` for current music tiers).
- **Example prompt:** “Upbeat lo-fi hip hop, no vocals, 90 BPM, warm Rhodes, 2 minutes.”

### get_music_status
- **Purpose:** Poll music job; mirror pattern of `get_generation_status`.

### generate_dialogue
- **Purpose:** Single-voice TTS. Supports two providers: **ElevenLabs v3** (default) and **Google Gemini 3.1 Flash TTS**. Both bill at 26 credits per 1,000 characters.
- **`provider`**: `"elevenlabs"` (default) or `"google"`.
- **ElevenLabs params:** `text` (or `prompt`, 5–5,000 chars after audio tags stripped), `voice` (one of 26 ids), `stability`, `similarity_boost`, `style`, `speed`, `previous_text`, `next_text`, `language_code` (ISO code, 29 supported).
- **Google params:** `text` (up to 4,000 bytes UTF-8; inline `[tags]` **performed**), `voice` (one of 30 voice ids; full list in the feature section above), `style_prompt` (natural-language delivery direction, up to 4,000 bytes; `text` + `style_prompt` combined must be under 8,000 bytes), `language_code` (BCP-47, e.g. `en-US`; or `auto`).
- **Model ids:** `text-to-dialogue-v3` (ElevenLabs) / `text-to-dialogue-gemini` (Google).
- **Example (ElevenLabs):** `{ "text": "Welcome back, ready to generate?", "voice": "Rachel" }`.
- **Example (Google):** `{ "provider": "google", "text": "[whispering] I have a secret. [laughing] Just kidding!", "voice": "Callirrhoe", "style_prompt": "Speak playfully, like sharing a fun secret." }`.
- **Multi-voice scenes:** call once per speaker, pass `previous_text` / `next_text` (ElevenLabs) for prosody continuity, then stitch the resulting audio.

### create_ad_copy
- **Purpose:** Variants of ad text for channels.
- **Key params:** `product_description`, `tone` (e.g. professional, playful), `platform` (e.g. Meta, TikTok, Google).

### get_models
- **Purpose:** Authoritative list of model ids, modality (video/image/music), and **credit cost hints**.
- **Use:** Call before `generate_media` / `generate_music` to validate ids.

### get_balance
- **Purpose:** Remaining **credits** for the authenticated account.

### get_generation_estimate
- **Purpose:** Estimate credit consumption before submitting a heavy job.

### get_upload_url / get_upload_session
- **Purpose:** Obtain a short-lived URL or session to **upload** a reference image/video for i2v or i2i pipelines when `source_image_url` is not already available.

### list_assets / add_asset / rename_asset / delete_asset (Asset Library)
- **Purpose:** Persistent **per-user named media library** — images / videos / audio the user wants to reuse across generations under a stable handle (e.g. `tesote-logo`, `brand-jingle`).
- **`list_assets()`** → returns each entry's `id`, `name`, `kind` (image|video|audio), `mime_type`, `size_bytes`, optional `width` / `height` / `duration_seconds`, plus a freshly signed CDN `url` valid for **1 hour** and `url_expires_at`. Also returns `quota_bytes` / `used_bytes`.
- **`add_asset(name, url)`** → server fetches the URL, validates MIME, stores it. `name` must match `^[a-z0-9_-]{1,64}$` and be unique per user. Free-tier quota: **50 MB total**, **500 MB per file**. Allowed types: jpeg / png / webp / mp4 / mov / mkv / webm / mp3 / wav / m4a / ogg / weba.
- **`rename_asset(asset_id, new_name)`** / **`delete_asset(asset_id)`** — manage the library; rename does not break past generations.
- **Reference in generations:** pass any asset's `url` field into `generate_media`'s `source_media_urls`. The library URL is a public, signed CDN link the model can fetch directly — no re-upload needed.

### Typical agent workflow
1. `get_models` → pick `model` + validate `generation_type` / `aspect_ratio` / `resolution` (for models that expose `resolution_options`).
2. Optional: `get_generation_estimate` → confirm credits.
3. If reference asset needed: check `list_assets` first — if the user's library already has it, pass the entry's `url` directly. Otherwise `get_upload_url` / `get_upload_session` (one-shot) **OR** `add_asset(name, url)` to save it for reuse.
4. `generate_media` → loop `get_generation_status` until `completed` → download `output_url`.
5. For music: `generate_music` → `get_music_status`.

---

### Discoverability URLs (new overview pages)

- Features hub: https://kubeez.com/features | RO: https://kubeez.com/ro/features
- Use cases: https://kubeez.com/use-cases | `/use-cases/social-media-creators` | `/use-cases/marketing-agencies` | `/use-cases/developers-api`
- Changelog: https://kubeez.com/changelog
- HTML sitemap: https://kubeez.com/sitemap