# Kuro

> An AI picture studio: prompt-driven images, clips, voice-overs and music, and
> storyboard-backed films with consistent characters across every scene. Work is
> account-wide, with no projects and no per-project asset libraries. Generation is
> priced in credits; every model below runs from the same composer.

## Agents

This product has an MCP server. A connected client can quote, generate, poll and download without a browser.

- Endpoint: https://api.meetkuro.com/mcp
- Install (Claude Code): `claude mcp add kuro https://api.meetkuro.com/mcp --transport http`

Authentication is OAuth; the tools dispatch through the app's own routes, so a client can do exactly what an account can do and nothing more.

- [Agent orientation](https://meetkuro.com/agents.md): one file for a program rather than a reader: the endpoint, how authentication works, what a run costs in credits, and the whole model and tool catalog. English only — it describes an interface whose tool names are English in every language.
- [Agents](https://meetkuro.com/agents/): the chapter for developers — what a connected client is taught, the tool names, and the raw JSON-RPC call shape in three runtimes.

## Guides

- [Grok Imagine Video 1.5: the complete prompting guide](https://meetkuro.com/guides/grok-imagine-video-1-5-prompting/): One picture in, up to fifteen seconds out, with sound you cannot turn off. Nine rules, three pipelines and the failures worth knowing about first.
- [P-Video: the complete guide to Pruna's four models](https://meetkuro.com/guides/p-video-prompting/): Four models, one family: a cheap generator and three performance rigs. What each one takes, what a take really costs, and the failures worth knowing first.
- [Seedance 2.5: the complete prompting guide](https://meetkuro.com/guides/seedance-2-5-prompting/): Thirty seconds of video with sound from one written brief. Six rules, three pipelines and the failures worth knowing about before you spend anything.
- [Veo 3.1: the complete prompting guide](https://meetkuro.com/guides/veo-3-1-prompting/): Eight seconds of video that writes its own soundtrack. Three audio channels, seven rules, four pipelines and the failures worth knowing about first.

## Tools

Single-purpose utilities. No prompt and no model to choose: each takes a file
and does the one thing it does.

- [Turn an image into an SVG you can actually edit](https://meetkuro.com/guides/image-to-svg/): Vectorize converts a logo, icon or flat illustration into real vector paths. What it is good at, what it will quietly ruin, and when to draw the vector instead.
- [Cut a subject out onto transparency, properly](https://meetkuro.com/guides/remove-a-background/): Remove background returns a lossless image with a real alpha channel. What it handles cleanly, what hair and glass do to it, and where the cut-out goes next.
- [Upscale an image without inventing detail](https://meetkuro.com/guides/upscale-an-image/): Crisp Upscale raises resolution and sharpens what is there. Where it belongs in a pipeline, why you generate small and enlarge last, and what it cannot recover.
- [Restore an old or damaged photograph](https://meetkuro.com/guides/restore-an-old-photo/): Restore photo repairs scratches, tears and faded colour with FLUX Kontext. What it recovers faithfully, what it reconstructs, and why that distinction matters.
- [Pull a frame out of a video, and keep going from it](https://meetkuro.com/guides/extract-a-frame-from-a-video/): Extract frame saves the first or last frame of a clip as a PNG, free. It is how you make a poster, a reference, and a longer film out of short takes.
- [Burn social captions into a video, automatically](https://meetkuro.com/guides/add-subtitles-to-a-video/): Add subtitles transcribes a clip and burns word-timed captions into it in one of four styles. The five-minute cap, which look to pick, and what it will not do.
- [Join short clips into one continuous film](https://meetkuro.com/guides/join-clips-into-one-video/): Join clips concatenates up to twenty of your own clips into one MP4, optionally under music and a voice-over. Free, agent-driven, and it appends nothing.

## How-tos

Questions asked across the catalog rather than about one entry in it.

- [Build an uncanny influencer that sells your product](https://meetkuro.com/guides/build-an-uncanny-influencer/): Invent a deadpan character from one line, lock her look in a bible, put her in a video she was never in, then hand her your product. Five steps.
- [Two pictures, fifteen seconds: the first-and-last-frame method](https://meetkuro.com/guides/first-and-last-frame/): Make the opening frame and the closing frame yourself, hand a video model both, and let it fill the middle. Most models take a last frame. What it buys.
- [Which image model to reach for, and when](https://meetkuro.com/guides/choosing-an-image-model/): Twelve image models, from half a credit to twenty-one. What each one is for, where the resolution setting changes the price, and when the cheap tier is right.

## Models

- [Seedream 5 Lite](https://meetkuro.com/models/seedream-5-lite/): The default image model: cheap, fast, takes seven references, and gives away its top resolution.
- [Nano Banana 2](https://meetkuro.com/models/nano-banana-2/): Google's image model, and the one to reach for when the aspect ratio is unusual.
- [GPT Image 2.5 Flare](https://meetkuro.com/models/gpt-image-2.5-flare/): OpenAI's instruction-follower, fast build. Six quality rungs, and the cheapest is a fortieth of the dearest.
- [GPT Image 2.5 Sunburst](https://meetkuro.com/models/gpt-image-2.5-sunburst/): The same model and the same price, tuned for the edit that has to change one thing and nothing else.
- [GPT Image 2](https://meetkuro.com/models/gpt-image-2/): The previous GPT Image, kept on the shelf for work already locked to it.
- [Nano Banana 2 Lite](https://meetkuro.com/models/nano-banana-2-lite/): Nano Banana's fourteen aspect ratios at half the price, capped at 1K.
- [Seedream 5 Pro](https://meetkuro.com/models/seedream-5-pro/): Seedream with more capability per pass, priced by size rather than given away.
- [FLUX Schnell](https://meetkuro.com/models/flux-schnell/): Half a credit an image. The volume tier, for when you need twenty of something.
- [FLUX.2 Flex](https://meetkuro.com/models/flux-2-flex/): FLUX.2 with more control per pass, priced a step above Pro.
- [FLUX.2 Pro](https://meetkuro.com/models/flux-2-pro/): Black Forest Labs' workhorse, and one of the cheapest reference-capable models here.
- [FLUX.2 Max](https://meetkuro.com/models/flux-2-max/): The top of the FLUX.2 ladder, for when the render itself is the deliverable.
- [Recraft V4.1 SVG](https://meetkuro.com/models/recraft-v4.1-svg/): The only way to get an editable vector out of this catalog. Logos, icons, flat illustration.
- [Recraft V4.1 Pro SVG](https://meetkuro.com/models/recraft-v4.1-pro-svg/): Recraft at 2048×2048: more path detail, at six times the price of the standard tier.
- [P-Image](https://meetkuro.com/models/p-image/): Half a credit, JPEG out. The other volume tier.
- [Grok Imagine Image](https://meetkuro.com/models/grok-imagine-image/): xAI's image model. Two credits, and one picture in for an edit rather than a reference set.
- [Grok Imagine Image Quality](https://meetkuro.com/models/grok-imagine-image-quality/): The sharper Grok: 2K output, better text rendering, and a resolution that is the whole decision.
- [Soul 2](https://meetkuro.com/models/soul-2/): Higgsfield's own photographic model. Half a credit at 720p or 1080p, and no references.
- [Seedance 2.5](https://meetkuro.com/models/seedance-2.5/): Thirty reference images, ten reference videos, ten reference audios, and takes up to thirty seconds.
- [Seedance 2.0](https://meetkuro.com/models/seedance-2.0/): The previous Seedance generation, and the only one in its family that renders 4K.
- [Seedance 2.0 Mini](https://meetkuro.com/models/seedance-2.0-mini/): The economical Seedance: same references and ceiling, capped at 720p.
- [Seedance 2.0 Fast](https://meetkuro.com/models/seedance-2.0-fast/): Seedance 2.0 tuned for turnaround rather than for price.
- [Grok Imagine Video 1.5](https://meetkuro.com/models/grok-imagine-video-1.5/): xAI's clip model. Needs an image to start from, and always generates audio.
- [Veo 3.1](https://meetkuro.com/models/veo-3.1/): Google's flagship: generates dialogue and sound from the same prompt as the picture.
- [Veo 3.1 Fast](https://meetkuro.com/models/veo-3.1-fast/): Veo without references, at roughly a third of the price. The single-shot tool.
- [Veo 3.1 Lite](https://meetkuro.com/models/veo-3.1-lite/): The volume tier of the Veo family. Audio is compulsory; 1080p only runs at eight seconds.
- [Gemini Omni 1.1](https://meetkuro.com/models/gemini-omni-1.1/): Google's other video model: sound with every take, and the only one here that edits a clip you already have.
- [Kling v3 Omni](https://meetkuro.com/models/kling-v3-omni-video/): Kuaishou's flagship: seven references, a reference video, and 4K output.
- [Kling v3](https://meetkuro.com/models/kling-v3-video/): Kling 3.0 without the reference channels: a prompt, a first and last frame, and 4K.
- [HappyHorse 1.0](https://meetkuro.com/models/happyhorse-1.0/): Alibaba's clip model. Prompt-only, five aspect ratios, up to fifteen seconds.
- [Wan 3.0](https://meetkuro.com/models/wan-3/): Alibaba's Wan 3.0. The one clip model here that runs a full thirty seconds at 1080p, from a prompt or one first frame.
- [Wan 3.0 Prime](https://meetkuro.com/models/wan-3-prime/): The same Wan 3.0 surface at roughly a third of the wait, for about 1.4 times the rate.
- [Runway Gen-4.5](https://meetkuro.com/models/gen-4.5/): Runway's model. 1080p only, five or ten seconds, six aspect ratios.
- [P-Video](https://meetkuro.com/models/p-video/): Pruna's general clip model. Cheap, flexible durations, one reference audio.
- [P-Video 2](https://meetkuro.com/models/p-video-2/): Pruna's quality-focused clip model. Up to twenty seconds, native speech, and first-to-last-frame control.
- [FLUX 3](https://meetkuro.com/models/flux-3/): Black Forest Labs' clip model. Picture and sound from one plain-language prompt, for up to twenty seconds.
- [MiniMax H3 Max](https://meetkuro.com/models/minimax-h3-max/): fal's own retuning of MiniMax H3. Fifteen seconds, native audio, and almost nothing to configure.
- [P-Video Avatar](https://meetkuro.com/models/p-video-avatar/): A photo plus an audio track becomes a talking clip. The prompt is optional.
- [P-Video Animate](https://meetkuro.com/models/p-video-animate/): Takes a still and a driving video, and moves the still the way the video moves.
- [P-Video Replace](https://meetkuro.com/models/p-video-replace/): Swaps the subject inside existing footage for one of your own.
- [OmniHuman 1.5](https://meetkuro.com/models/omni-human-1.5/): ByteDance's performance model: a portrait and an audio track, at 1080p with real body motion.
- [Kling Avatar v2](https://meetkuro.com/models/kling-avatar-v2/): A portrait and a voice track become a talking clip — people, animals or cartoons alike, at 720p.
- [Kling Motion Control](https://meetkuro.com/models/kling-v3-motion-control/): Drives a still image with the motion of a video. Prompt optional, 720p or 1080p.
- [Kling Lip Sync](https://meetkuro.com/models/kling-lip-sync/): Six credits: re-syncs a person in existing footage to a new audio track.
- [Genjutsu Motion Transfer](https://meetkuro.com/models/genjutsu-motion-transfer/): Higgsfield's motion transfer: up to eight images of your subject, performing a source video's motion.
- [Genjutsu Object Swap](https://meetkuro.com/models/genjutsu-object-swap/): Higgsfield's object swap: the source video stays, and the thing in it becomes the one in your images.
- [Eleven v3](https://meetkuro.com/models/elevenlabs-tts/): Expressive voice-over directed by inline audio tags, with twenty-six preset voices and saved ElevenLabs voices.
- [Eleven v4](https://meetkuro.com/models/eleven-v4/): Expressive voice-over with inline audio tags, George, and your saved ElevenLabs voices.
- [Gemini 3.1 Flash TTS](https://meetkuro.com/models/gemini-3.1-flash-tts/): Thirty voices across twenty-five languages, from Google.
- [Kokoro 82M](https://meetkuro.com/models/kokoro-82m/): Thirty-three English-accent voices, priced for volume.
- [MiniMax Speech 2.8 HD](https://meetkuro.com/models/minimax-speech-2.8-hd/): The directable one: emotion selector, speed control, twenty-four languages.
- [Inworld TTS 2.0](https://meetkuro.com/models/inworld-tts-2/): Four voices, sixteen languages, tuned for realtime delivery.
- [MiniMax Music 2.6](https://meetkuro.com/models/minimax-music-2.6/): Prompt-to-track, and the one music model here that sings.
- [Eleven Music](https://meetkuro.com/models/eleven-music/): A minute of instrumental scoring per take, from a model trained on licensed music.

## Comparisons

- [Seedream 5 Lite vs Nano Banana 2](https://meetkuro.com/compare/seedream-5-lite-vs-nano-banana-2/): Take Seedream 5 Lite unless the deliverable has an unusual shape. Then it has to be Nano Banana 2.
- [Nano Banana 2 vs GPT Image 2](https://meetkuro.com/compare/nano-banana-2-vs-gpt-image-2/): GPT Image 2 when the prompt is a set of instructions; Nano Banana 2 when the shape is unusual or the volume is high.
- [Veo 3.1 vs Seedance 2.5](https://meetkuro.com/compare/veo-3.1-vs-seedance-2.5/): Veo 3.1 when someone speaks. Seedance 2.5 when the shot is long, or has a cast to keep consistent.
- [Veo 3.1 vs Veo 3.1 Fast](https://meetkuro.com/compare/veo-3.1-vs-veo-3.1-fast/): Fast for a shot that stands alone. Full Veo the moment the clip has to match something else.
- [P-Video Avatar vs OmniHuman 1.5](https://meetkuro.com/compare/p-video-avatar-vs-omni-human-1.5/): Avatar for volume and for faces in the background. OmniHuman when the person is the subject.
- [FLUX.2 Pro vs FLUX.2 Max](https://meetkuro.com/compare/flux-2-pro-vs-flux-2-max/): Iterate on Pro, render the keeper on Max, and only when the image is the finished deliverable.

## Arena

A weekly ranking of image and video models on published prompts, with every output and its real credit cost shown side by side. Editorially judged; the standings are recomputed from the full matchup history on every build.

- [The arena](https://meetkuro.com/arena/): the boards, the method, and every published week.
- [Standings](https://meetkuro.com/arena/leaderboard.json): the current leaderboard as JSON, CC BY 4.0 — reuse it with a link.
- [Arena week 36: every model can spell now](https://meetkuro.com/arena/2026-w36/): Five image models, four prompts, one week. All five rendered accented French correctly — so the ranking came down to hands, fabric and whether 'exactly three' means three.

## About

- [Model catalog](https://meetkuro.com/models/): every model, grouped by kind, with what each costs in credits.
- [Guides](https://meetkuro.com/guides/): the full index.
- [Tools](https://meetkuro.com/guides/#tool): the finishing tools, one guide each.
- [Pricing](https://meetkuro.com/pricing/): plans and monthly credit allowances.
- [Terms](https://meetkuro.com/terms/) · [Privacy](https://meetkuro.com/privacy/)

The app itself is at https://app.meetkuro.com and requires an account; it is
disallowed in that origin's robots.txt because every route behind it is a
signed-in surface.