app.nzapp
AppsProjectsReposPullsChatIntegrationsGatewayModelsEvalsToolsDatasetsMCPDeploysPricingBlogDocsAssistantsCharactersArtMusic
Sign inStart building
Agent stack
Cloud coding agentAgents SDKIntegrationsBrowser agentMonitors & auto-agentsSchedulersAgent skillsMCP serversDeep research
Models & API
AI GatewayModel catalogModel evalsModel spacesPlaygroundText to imageImage to 3DText to 3DMusic & SFXAudio editorMedia optimizerAI art & libraryChatAPI referenceSchemaBecome a provider
Compute & hosting
DeploysAddonsPostgres hostinggobed vector searchSite hostingAnalyticsCog GPU hostingRL trainingBuilds & CIWorkersTask queuesDomainsGit hosting
Tools
AI toolsDrawDiffusion canvasLive DrawWriteSheetsArtifactsVideo studioNotebooksDatasets
Learn
DocsBlogEval guidesPrompt libraryCLIAlternativesPapersAI charactersArt gallerySecurityConsulting
Company
PricingEnterpriseSettingsBillingStatusInvestorsCreate accountTerms of ServicePrivacy Policy
app.nzapp.nz

AI agent cloud for coding, deploys, model routing, and research. Built for teams shipping software.

Built in New Zealand by App AI NZ.

Social
X / TwitterGitHubYouTube
The app.nz network
GpuBrainPapersReading TimemojojojoNetwrckText-Generator.ioCodex InfinityOpenPathsCuteDSLAI Art GeneratorAIArt-Generator.artSiteSimSimplexGenDictatorFlowWebFiddleRing.nzChatGibidyBitBankExperimentFlowEvangelerHires.nzHow.nzV5 GamesAddicting Word GamesBig Multiplayer ChessWord SmashingreWord GameMultiplication Master
© 2026 App AI NZ Ltd. All rights reserved.All systems normalTermsPrivacy
Blog
August 4, 2026·8 min read·app.nz

Packaging MiniMax H3 for Cog, R2, RTX 5090, and scale to zero

A careful H3 video workflow with text, keyframes, native audio, true loops, verified R2 weights, GPU AV1, serverless pricing, and a fail-closed license gate.

Listen to this article

On-device voice

Uses the voice built into your browser; no article text leaves this page.

Audio narration is not supported by this browser.

MiniMax H3 text, keyframe, audio, and loop workflow
MiniMax H3 text, keyframe, audio, and loop workflow

MiniMax H3 is an unusually complete open-weight video pipeline: text-to-video, image-to-video, first/last-frame control, and synchronized stereo audio share one model. We packaged that surface as a portable Cog and a RunPod Serverless worker, then added the operational pieces a hosted model needs: verified R2 weights, scale-to-zero routing, safe media inputs, and GPU AV1 delivery.

The adapter source is github.com/lee101/h3-cog. Its MIT license covers our code, not the model weights.

One workflow, four useful modes

The same endpoint accepts a prompt plus zero, one, or two keyframes:

modeinputuseful for
text-to-videopromptideation and complete audiovisual shots
image-to-videoprompt + first frameanimating art, products, and storyboards
first/last frameprompt + both framestransitions with a controlled landing
loopprompt + first frame + loopa generated clip whose final condition is the opening frame

Loop mode is not a reversed copy or cross-fade. The first frame is also passed to H3 as its final-frame condition, so the model generates the return motion. That gives the sampler a real chance to preserve both the subject and the seam.

H3 renders at 24 fps. Duration is snapped to its valid 17k+5 frame grid; a five-second request becomes 124 frames, or about 5.17 seconds. Native 16:9 is 1344×768. The Cog also exposes balanced and preview canvases for cheaper prompt iteration before a final native render.

The 5090 path

Our selected ComfyUI checkpoint set is about 42.5GB: a 20.97GB pruned INT8 diffusion model, 15.69GB NVFP4 text/vision encoder, 5.21GB video VAE, and 0.61GB audio VAE. That is larger than a 32GB RTX 5090, so the runtime deliberately offloads between text encoding, denoising, and VAE decode instead of pretending every component remains resident.

We pin the ComfyUI commit that introduced the native H3 nodes, use the official res_multistep sampler at 20 steps, and enable SageAttention when available. The current H3 release is dense full-attention. We did not transplant a CG-Taylor-style cache without quality evidence. A separate acceleration note explains the signed EasyCache A/B path now available to operators; the public default remains uncached until it passes joint audio/video and first/last-frame quality gates.

The output stage prefers av1_nvenc on Blackwell and preserves generated audio as Opus in WebM. It falls back to SVT-AV1 when NVENC is unavailable; H.264/AAC is available for clients that value compatibility over size.

R2 weights without mystery files

Weights are runtime data, not Docker layers. The mirror tool downloads only the four production files, calculates SHA-256, uploads immutable objects to R2, and publishes a small manifest. A worker resumes partial HTTP downloads, verifies both size and digest, and only then links a file into ComfyUI. A shared RunPod volume means a cold process does not imply a 42.5GB cold download.

This split keeps the open-source container reviewable and gives app.nz a fast, repeatable cache without weakening provenance.

Pricing and scale to zero

The catalog template targets a 32GB RTX 5090, keeps no minimum workers, and has a five-minute warm window for bursts. Its displayed serverless estimate is the selected RunPod compute rate plus an 80% app.nz platform fee, charged by execution second. Idle is zero.

The template exposes the same schema through the website, Cog HTTP API, CLI, desktop proxy, Comfy/MCP template discovery, and RunPod handler. There is no second, quietly divergent inference contract.

Why launch is paused today

MiniMax's current model license defines an applicable territory that excludes the United States, European Union, United Kingdom, and Republic of Korea. The existing app.nz Cog pod path is pinned to a US region, and its legacy RunPod Serverless integration does not yet prove a compliant data-center restriction.

So the catalog is fail-closed: it shows the workflow, source, schema, warning, and final pricing, but the launch button cannot allocate a worker. The runtime also refuses to download weights unless the deployer explicitly sets MINIMAX_H3_LICENSE_ACCEPTED=1. We will not accept model terms on a user's behalf or move restricted weights into R2 first and ask questions later.

Once acceptance and a permitted RunPod location are supplied, the release gate is straightforward: mirror the four files, build the pinned container, run the seven deterministic launch prompts (text, first-frame, first/last-frame, true loop, vertical, square, and ultrawide), visually review first/last-frame identity and audio sync, publish the clips, then enable the one-click button.

Review the H3 Cog source, inspect the upstream model card, or open the H3 catalog preview.

Build what you just read

Ship agents, models, and apps on one cloud.

Start with free credits, then use the same platform from the web app, CLI, desktop app, or MCP.

Start building freeRead the docs

Keep reading

Train three open piano models, then deploy one Cog anywhere

A reproducible CC0 symbolic training pipeline, listening previews, one inference image for Cog and RunPod Serverless, and one-click scale-to-zero deployment on app.nz.

From ComfyUI to any Cog: open audio workflows that scale to zero

How template:name connects ComfyUI to reusable Cog deployments, when to choose open Pocket TTS versus fal audio, and how the shared serverless/pod router returns idle workers to zero.

Packaging AlayaWorld: an interactive video world model as a self-hosted Cog

Wrapping AlayaLab’s autoregressive world model (LTX-2.3-derived DiT + gemma-3-12b + Depth-Anything-3) as a Cog: camera presets, a warm engine, and an honest look at the LTX-2 Community License.