Veo 3.1 Fast MCP guide for premium campaign video

Preflight one coherent campaign beat

Reduce the concept to one observable beat: one subject, one setting, one primary action, one camera move, and one lighting transition. For example: “A brushed-steel perfume bottle rotates a quarter turn on black stone while a narrow amber light travels across the label, slow dolly-in, restrained haze, no added text.” This is an editorial prompt pattern, not evidence that the model will preserve a logo, render exact lettering, or follow every instruction.

Write acceptance criteria before generation: the product silhouette remains recognizable, the intended label region is not replaced with misleading copy, the camera move does not hide the product, no unrequested people or marks appear, and the final framing works in the selected campaign placement. Review the completed output frame by frame. A “premium campaigns” catalog label describes intended fit; it does not replace legal, brand, accessibility, or visual-quality review.

Keep Studio controls separate from fal capabilities

fal's direct Veo 3.1 Fast documentation identifies its endpoint as `fal-ai/veo3.1/fast`. The schema requires a prompt, offers 16:9 and 9:16 aspect ratios, accepts `4s`, `6s`, or `8s`, and defaults to 16:9 and `8s`. It also documents resolution, audio, seed, negative-prompt, automatic-fix, and safety controls. The cited Studio sources do not establish that this exact endpoint identifier is Studio's configured route.

Published `@offlinecreator/mcp` 0.1.2 gives `generate` four inputs: `modelId`, `prompt`, optional `aspectRatio`, and optional `wait`. Its package-wide ratio enum includes 16:9, 9:16, and 1:1, and its description says the frame defaults to 16:9. It has no duration, resolution, audio, seed, negative-prompt, automatic-fix, or safety-tolerance argument. The public Studio catalog does not publish Veo-specific ratio compatibility, so this draft leaves model-specific Studio ratios unknown rather than treating the package-wide enum as proof.

Ratios shared by the two cited schemas

16:9 or 9:16This compares published schemas; it does not prove Studio's endpoint mapping.

Provider duration values

4s, 6s, or 8s; default 8sProvider schema only; published MCP 0.1.2 has no duration input.

Published MCP inputs

modelId, prompt, aspectRatio, waitDo not infer direct-provider controls from fal documentation.

Approve the 96-credit attempt before submission

The current Studio catalog prices Veo 3.1 Fast at 96 credits, 54 credits more than the listed 42-credit cost of either launch Kling video path, or about 2.29 times as much. That arithmetic describes today's catalog only; it is not a benchmark showing that Veo is better, faster, more reliable, or more suitable for every campaign. OfflineCreator states that web, MCP, CLI, and API generations draw from the same account balance and published per-model costs.

Before one paid attempt, record the current `list_models` response, available credits, prompt, acceptance criteria, and operator approval. Submit one clip, review it, and decide whether another run is justified. Do not convert fal's direct per-second pricing into Studio credits: the Studio catalog already supplies the customer cost, and fal's billing options include variables that the published MCP tool does not expose.

Use Veo 3.1 Fast for one bounded premium-campaign shot

OfflineCreator Studio currently lists Veo 3.1 Fast as a text-to-video model for premium campaigns at 96 credits per generation. In MCP, call `list_models` immediately before the run, select the Veo entry from that response, and pass its returned Studio model ID to `generate`. The cited public catalog displays the model, workflow, positioning, and cost, but its visible catalog does not display a Studio model-ID field.

Treat eight seconds as a direct-provider default documented by fal, not as a tested Studio duration or a quality signal. fal's separately published Veo 3.1 Fast endpoint accepts four-, six-, and eight-second values and defaults to eight seconds. Published OfflineCreator MCP 0.1.2 exposes no duration argument, so this page does not claim which duration Studio submits or that an MCP caller can choose any provider duration.

Current Studio listing

Veo 3.1 Fast · text-to-video · 96 creditsPositioned for premium campaigns; refresh the live catalog before spending.

Studio MCP model ID

Use the value returned by list_modelsThe accessible public catalog does not publish a stable ID for copying into automation.

Provider endpoint ID

fal-ai/veo3.1/fastThis identifies fal's endpoint; it is not a substitute for the Studio model ID.

Run MCP discovery before generation

Use `list_models` first because the published MCP server describes it as the launch-catalog tool for model workflows and credit costs. Then call `generate` with the Studio ID returned for the selected Veo entry and the approved prompt. Published package 0.1.2 offers a package-wide `aspectRatio` enum of 16:9, 9:16, and 1:1, but the cited public Studio catalog does not establish which of those values Veo accepts; model-specific Studio ratio compatibility remains unknown. In the same package, `wait: true` polls a text-to-video job for up to 120 seconds, while `get_generation` returns current status and `wait_generation` performs bounded polling.

Do not paste `fal-ai/veo3.1/fast` into `generate.modelId` on the strength of fal documentation alone. The MCP package defines `generate.modelId` as an ID from `list_models`; fal identifies its string as a direct-provider endpoint. Keeping them separate avoids claiming an unverified Studio endpoint mapping and makes the live Studio catalog response the operational source for its own identifier.

Choose the next page by the unresolved decision

Return to the model directory to refresh catalog availability, use the video-credit guide to set a broader experiment budget, or open the identifier guide when an integration must distinguish Studio IDs from provider endpoints. These links answer the next operational question without expanding this page into a cross-model benchmark.

Required evidence and publication boundary

This page owns only “veo 3.1 fast mcp.” Current public Studio evidence supports the Veo 3.1 Fast listing, text-to-video workflow, premium-campaign positioning, and 96-credit cost. The public catalog's visible fields do not include a Studio model ID, while the published MCP package directs callers to use an ID returned by `list_models`. The required public Studio-ID evidence is therefore incomplete, and this page does not substitute fal's direct endpoint identifier.

fal's direct documentation supports its own endpoint ID, duration values, aspect ratios, and provider controls. Published OfflineCreator MCP 0.1.2 supports the narrower MCP input surface and exposes no duration, resolution, or audio input. Those sources do not prove which fal endpoint or duration Studio currently submits. This research performed no paid generation or output-quality test and uses no community claim. Reddit retrieval was partial, Digg was rate-limited, and X was unavailable, so the page remains a draft, unreviewed, and unpublished.