Google's programmatic endpoint for Veo 3.1 — generate short video with synchronized native audio from code, across three price tiers (Standard, Fast, and Lite), with reference-image consistency and scene extension.
Last verified · 2026-10-17 · by Moe Ameen
The Gemini Veo 3.1 API is the programmatic way to use Google DeepMind's Veo 3.1 video model, which launched on October 15, 2025. Instead of generating clips by hand in the Gemini app or Google Flow, you call an endpoint with a prompt and get back a short video with synchronized native audio. The API exposes three variants, each with its own preview model ID: Veo 3.1 Standard (`veo-3.1-generate-preview`), the highest-fidelity tier; Veo 3.1 Fast (`veo-3.1-fast-generate-preview`), a cheaper tier for iteration; and Veo 3.1 Lite (`veo-3.1-lite-generate-preview`), the lowest-cost tier. It is reachable through the official google-genai Python and JavaScript SDKs and a REST endpoint, and the same models are also available on Vertex AI for Google Cloud teams.
Technically, clips run 4, 6, or 8 seconds at 24fps, in 16:9 or 9:16, with native audio — dialogue, sound effects, and ambient sound, lip-synced to speaking characters — generated in the same pass. Resolution defaults to 720p, with 1080p and (on Standard and Fast) 4K available only on 8-second clips. Standard and Fast accept up to three reference images to hold a character, product, or style consistent, and support scene extension that chains clips by adding 7 seconds at a time up to roughly 148 seconds; Lite omits reference images, scene extension, and 4K. Generation is asynchronous: the call returns a long-running operation you poll until it finishes, then you download the file.
Pricing is per second of output on the paid tier. At the time of writing, Standard runs about $0.40/second at 720p and 1080p and $0.60 at 4K; Fast is roughly $0.10–$0.12 at 720p/1080p and $0.30 at 4K; and Lite is $0.05–$0.08 with no 4K — so an 8-second Standard clip is near $3.20. There is no free tier, and Google bills only successful generations. Because all three endpoints are preview models and the earlier Veo 3 IDs are already deprecated, treat the specifics above as the current shape rather than fixed guarantees: Google's own Gemini API deprecations page already lists October 22, 2026 as the earliest possible shutdown date for all three Veo 3.1 IDs, naming `gemini-omni-1.1-flash` as the recommended replacement — confirm model IDs, limits, and per-second prices on Google's own pages before building on them.
The Veo 3.1 API's most underrated feature for a brand is reference-image consistency: pass the same character or product as a reference across calls and you get clips that look like one continuous identity instead of a new face every render. That is exactly the raw material a recurring content series needs — but the API still only ever hands you an 8-second clip with no caption track, no brand frame, and only the one aspect ratio you asked for. Kompozy is where a consistent character becomes a recurring, on-brand, published format.
Generate a run of reference-consistent clips through the API, drop them into Kompozy, and Clipped Shorts cuts each one to vertical 9:16 with word-synced captions (the track every clip needs because feeds autoplay muted, even with Veo's native audio). Because the visual identity is already locked by your reference image, Kompozy layers the brand on top: a Carousel built with pixel-exact HyperFrames, Quote Graphics of the key line, a Photo Post, a Blog Article, and an Email Newsletter — all governed by your Persona Brief so the copy stays on-voice while Veo keeps the character on-model. Kompozy's own face-locked persona pool then carries that identity into recurring avatar Persona Shorts between API renders, so the series never breaks character. Autopilot and a per-post review pipeline schedule and publish the whole set across eight social platforms plus blog and email. The API holds the character for eight seconds; Kompozy turns it into a weekly branded show.
It is Google's programmatic endpoint for the Veo 3.1 video model (launched October 15, 2025), reachable via the google-genai Python/JavaScript SDKs and REST. You call it with a prompt — and optionally a starting image, up to three reference images, or a last frame — and it generates a short clip with synchronized native audio, which you retrieve by polling a long-running job.
Three preview endpoints: `veo-3.1-generate-preview` (Standard, highest fidelity), `veo-3.1-fast-generate-preview` (Fast, cheaper for iteration), and `veo-3.1-lite-generate-preview` (Lite, lowest cost, no 4K or reference images). The older Veo 3 IDs are deprecated, and Google's own deprecations page already lists October 22, 2026 as the earliest possible shutdown date for all three Veo 3.1 IDs too, pointing to `gemini-omni-1.1-flash` as the replacement. Pin the exact ID, since preview models can be retired.
Per second of output on the paid tier: at the time of writing, Standard is about $0.40/second at 720p and 1080p ($0.60 at 4K), Fast about $0.10–$0.12 ($0.30 at 4K), and Lite $0.05–$0.08 with no 4K. An 8-second Standard clip is roughly $3.20. There is no free tier, and only successful generations are billed. Confirm live rates before budgeting.
A single generation is 4, 6, or 8 seconds (the 8-second length is required for 1080p, 4K, reference images, and extension). Scene extension on Standard and Fast adds 7 seconds per call, up to 20 times, reaching about 148 seconds — but only on a Veo-generated 720p source, and the extended output stays 720p.
Kompozy is the finishing and distribution layer. Generate reference-consistent clips through the API, bring them into Kompozy, and it cuts vertical captioned shorts, reframes per feed, and spins the same identity into carousels, quote graphics, a blog, and a newsletter in your Persona Brief voice — then publishes across nine destinations: eight social platforms plus blog and email.