OfflineCreator MCP model guide directory
Route users by workflow, outcome, speed, detail, motion, and credit budget.
Start with the output you need: a still image, a video from text, or motion from an existing image. Then call `list_models` in OfflineCreator MCP before `generate`. The published package describes `list_models` as the launch catalog with credit costs, while `generate` expects a model ID from that catalog.
The current catalog has six launch models across text-to-image, text-to-video, and image-to-video. Treat this as a dated directory, not a permanent compatibility promise: model names, costs, and supported inputs can change, so refresh the catalog in the client instead of copying an old identifier.
- npm Registry / OfflineCreator Studio: @offlinecreator/mcp 0.1.1
- OfflineCreator Studio via unpkg: @offlinecreator/mcp 0.1.1 README
- OfflineCreator Studio via unpkg: @offlinecreator/mcp 0.1.1 published server implementation
- OfflineCreator Studio: Models and credit costs
- OfflineCreator Studio: MCP and CLI for AI image and video generation
Current launch catalog
For still images, the catalog positions FLUX Schnell for fast iteration, FLUX 1.1 Pro Ultra for photoreal detail, and Recraft V3 for brand graphics. For video, Kling 2.6 Pro is the text-to-video option, Kling 2.6 Pro Motion animates an image, and Veo 3.1 Fast is the premium text-to-video option.
All six current launch entries route through fal. That provider label describes Studio's present routing and does not mean every underlying model is made by fal.
- FLUX Schnell
- FLUX Schnell · text-to-image · 1 creditCurrent first-party label: fast iteration.
- FLUX 1.1 Pro Ultra
- FLUX 1.1 Pro Ultra · text-to-image · 8 creditsCurrent first-party label: photoreal detail.
- Recraft V3
- Recraft V3 · text-to-image · 5 creditsCurrent first-party label: brand graphics.
- Kling 2.6 Pro
- Kling 2.6 Pro · text-to-video · 42 creditsCurrent first-party label: social video.
- Kling 2.6 Pro Motion
- Kling 2.6 Pro Motion · image-to-video · 42 creditsCurrent first-party label: image animation.
- Veo 3.1 Fast
- Veo 3.1 Fast · text-to-video · 96 creditsCurrent first-party label: premium campaigns.
A practical MCP decision path
For a low-cost concept pass, start with FLUX Schnell. Move to FLUX 1.1 Pro Ultra when the brief specifically calls for the catalog's photoreal-detail fit, or use Recraft V3 when graphic composition is the priority. Those labels are product positioning, not comparative quality benchmarks.
For motion, decide whether the first frame already exists. Choose the image-to-video workflow only when you have an image to attach: the published MCP server reserves the generation first, then `upload_input` attaches the image and submits it. Choose a text-to-video entry when the shot should begin from a written description.
A useful request states the subject, setting, composition, light, and intended frame. The published `generate` schema accepts 16:9, 9:16, or 1:1, but clients should still call `list_models` and use the current model data rather than infer model-specific support from the tool's broad input schema.
Budget the shortlist before generation
OfflineCreator publishes model-specific credit costs and says MCP, CLI, API, and web Studio generations use the same account balance. The cost is shown before generation. These are current product credits, not a guarantee about runtime, visual quality, or future pricing.
For example, one FLUX Schnell draft plus one FLUX 1.1 Pro Ultra generation uses 9 credits at the current catalog rates. Three Recraft V3 variations use 15 credits. This arithmetic is a planning example only; it does not predict how many attempts a creative task will require.
Continue with the model-specific guide
Open the narrower guide for the model or budgeting question you have next, or return to the model hub for the maintained catalog.
Directory boundary
This page owns model routing for the query “offlinecreator mcp models.” Keep detailed prompting, output evaluation, and model-specific caveats on their narrower guides. Consolidate this directory into the model hub if the catalog source can no longer be maintained.
The complete evidence refresh retained 33 item-level records: 12 Digg clusters, 11 Reddit threads, one TikTok video, and nine YouTube videos. Review classified two as relevant, eight as marginal, 13 as irrelevant, four as duplicates, five as promotional, and one as unverifiable. The relevant discussions concerned prompt rewriting in Veo and practitioner reactions to other image or video models; neither verified OfflineCreator's launch catalog, provider routing, costs, or supported inputs, so no source, claim, or citation was added.
Coverage remains degraded. Reddit returned 11 records before HTTP 429, Instagram failed with HTTP 404, and X was unavailable because optional credentials were not configured. All nine YouTube records lacked publication dates in the raw export and reported zero captured views, likes, and comments. Their generic MCP introductions and competitor setup guides were not treated as current OfflineCreator model evidence.
Recent model anecdotes can suggest questions for a matched-output test, but they cannot establish OfflineCreator compatibility, speed, visual quality, reliability, pricing, or outcomes. The version-pinned package artifacts and first-party catalog remain the stronger evidence for this directory. No fresh generation run, prompt-rewrite inspection, or matched-output evaluation was performed.