Veo 3 API

Integrate Google Veo 3.1 Fast video synthesis across text-to-video, image-to-video, reference-to-video, and start-end-frame-to-video. Text, image, and reference endpoints bill $0.01/s silent or $0.015/s default with synchronized audio at 720p/1080p, up to $0.0225/s in 4K.

Catalog scope note: E2X currently serves google/veo-3.1-fast (text-to-video, image-to-video, reference-to-video, start-end-frame-to-video). Standard non-fast veo-3 and veo-3.1 variants are not in the E2X catalog.

Endpoints Covered
4
Lowest Catalog Rate
$0.01/s
Capabilities
Start End Frame To Video, Image To Video, Reference To Video, Text To Video
Savings vs fal.ai
Up to 90%

01 / LIVE CATALOG PRICING

Real Catalog Pricing & Dynamic Tiers

Rates below are calculated dynamically from E2X micro-cent pricing rules and default resolution/quality multipliers.

Model & Endpoint SlugCapabilityDefault RateDynamic Pricing Tiers
Veo 3.1 Fast

google/veo-3.1-fast/start-end-frame-to-video

Start End Frame To Video$0.10/s
Fixed rate (/s):$0.10/s
Veo 3.1 Fast

google/veo-3.1-fast/image-to-video

Image To Video$0.015/s
720p · Silent:$0.01/s720p · With sound (default):$0.015/s1080p · Silent:$0.01/s1080p · With sound:$0.015/s4k · Silent:$0.015/s4k · With sound:$0.0225/s
Veo 3.1 Fast

google/veo-3.1-fast/reference-to-video

Reference To Video$0.015/s
720p · Silent:$0.01/s720p · With sound (default):$0.015/s1080p · Silent:$0.01/s1080p · With sound:$0.015/s4k · Silent:$0.015/s4k · With sound:$0.0225/s
Veo 3.1 Fast

google/veo-3.1-fast/text-to-video

Text To Video$0.015/s
720p · Silent:$0.01/s720p · With sound (default):$0.015/s1080p · Silent:$0.01/s1080p · With sound:$0.015/s4k · Silent:$0.015/s4k · With sound:$0.0225/s
Model job total with batch counts, resolution, and duration before spending credits.Estimate costs in the Pricing Calculator

02 / QUICKSTART CODE

SSR Code Samples (cURL, JavaScript, Python)

Submit requests directly against POST https://api.e2x.ai/v1/jobs/submit or use the official @e2x/sdk and Python e2x clients with "google/veo-3.1-fast/text-to-video".

cURLPOST google/veo-3.1-fast/text-to-video
curl -sS -X POST "https://api.e2x.ai/v1/jobs/submit" \
  -H "Authorization: Bearer $E2X_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "google/veo-3.1-fast/text-to-video",
  "input": {
    "prompt": "Slow dolly push through a misty pine forest at dawn with warm sunbeams cutting across the trail",
    "aspect_ratio": "16:9",
    "resolution": "720p",
    "duration": "8",
    "has_sound": "true"
  }
}'
JavaScriptNode.js / JavaScript (@e2x/sdk)
import { E2XClient } from '@e2x/sdk';

const e2x = new E2XClient({ apiKey: process.env.E2X_API_KEY });

const result = await e2x.jobs.run({
  model: 'google/veo-3.1-fast/text-to-video',
  input: {
    "prompt": "Slow dolly push through a misty pine forest at dawn with warm sunbeams cutting across the trail",
    "aspect_ratio": "16:9",
    "resolution": "720p",
    "duration": "8",
    "has_sound": "true"
  },
});

console.log(result.output);
PythonPython (e2x SDK)
import os
from e2x import E2XClient

client = E2XClient(api_key=os.environ["E2X_API_KEY"])

job = client.jobs.run(
    model="google/veo-3.1-fast/text-to-video",
    input={
        "prompt": "Slow dolly push through a misty pine forest at dawn with warm sunbeams cutting across the trail",
        "aspect_ratio": "16:9",
        "resolution": "720p",
        "duration": "8",
        "has_sound": "true"
    },
)

print(job.output)

03 / API RESPONSES & WEBHOOKS

Real API Response Payloads & Error Envelope

Submission, polling, HMAC-signed webhook callbacks, and error envelopes share one consistent JSON structure across all E2X endpoints.

POST /v1/jobs/submit (Queued Response)

JSON200 OK
{
  "success": true,
  "data": {
    "jobId": "job_01J9XK2M8N4P7Q3R5S6T8V0WXY",
    "status": "pending",
    "estimatedCompletionTime": 35,
    "costMicroCents": 120000
  },
  "requestId": "req_01J9XK2M8N4P7Q3R"
}

GET /v1/jobs/:jobId (Completed Output)

JSON200 OK
{
  "success": true,
  "data": {
    "jobId": "job_01J9XK2M8N4P7Q3R5S6T8V0WXY",
    "model": "google/veo-3.1-fast/text-to-video",
    "status": "completed",
    "progress": 100,
    "output": "https://assets.e2x.ai/generated/example.mp4",
    "outputs": [
      {
        "url": "https://assets.e2x.ai/generated/example.mp4",
        "type": "video"
      }
    ],
    "failedGenerations": [],
    "numberOfGenerations": 1,
    "requestedCount": 1,
    "successCount": 1,
    "failedCount": 0,
    "costMicroCents": 120000,
    "refundedMicroCents": 0,
    "error": null,
    "createdAt": "2026-11-13T10:00:00.000Z",
    "completedAt": "2026-11-13T10:00:18.000Z"
  },
  "requestId": "req_01J9XK2N1B2C3D4E"
}

Webhook Callback Body (X-E2X-Signature HMAC-SHA256)

JSONPOST webhookUrl
{
  "jobId": "job_01J9XK2M8N4P7Q3R5S6T8V0WXY",
  "requestId": "req_01J9XK2M8N4P7Q3R",
  "status": "completed",
  "output": "https://assets.e2x.ai/generated/example.mp4",
  "error": null,
  "completedAt": "2026-11-13T10:00:18.000Z"
}

Verify X-E2X-Signature (HMAC-SHA256 over <X-E2X-Timestamp>.<rawBody>). Read Webhook Signature & Retry docs →

Standard Error Envelope

JSON4xx / 5xx JSON Error
{
  "success": false,
  "error": {
    "code": "CONCURRENT_JOBS_EXCEEDED",
    "message": "Concurrent job limit reached for this model."
  },
  "requestId": "req_01J9XK2Q9W8E7R6T"
}

Every non-2xx response returns success: false with error.code and error.message. Browse API error codes →

04 / INPUT SCHEMA

Input Parameter Reference

Parameters and constraints extracted directly from the live endpoint JSON schemas for covered models.

ParameterTypeRequiredDefaultAllowed Values / ConstraintsEndpointsDescription
promptstringRequired—maxLength: 50000All 4 endpoints—
aspect_ratiostringRequired"16:9"enum: 16:9, 9:16All 4 endpoints—
seedintegerOptional——veo-3.1-fast/text-to-video, veo-3.1-fast/image-to-video, veo-3.1-fast/reference-to-video—
resolutionstringOptional"720p"enum: 720p, 1080p, 4kAll 4 endpointsResolution
durationstringOptional"8"enum: 4, 6, 8All 4 endpointsDuration
has_soundstringOptional"true"enum: false, trueveo-3.1-fast/text-to-video, veo-3.1-fast/image-to-video, veo-3.1-fast/reference-to-videoEnable Sound
start_frame_urlstringRequired——veo-3.1-fast/start-end-frame-to-videoStart frame image URL
end_frame_urlstringRequired——veo-3.1-fast/start-end-frame-to-videoEnd frame image URL
image_urlstringRequired"https://assets.e2x.ai/uploads/0a2491b4-a2c1-444c-ae7c-95c6e9b0db38.png"—veo-3.1-fast/image-to-video—
image_urlsarrayRequired—items: 1..unboundedveo-3.1-fast/reference-to-video—

05 / VERIFIED SAMPLE OUTPUTS

Catalog Sample Outputs

Example media served from assets.e2x.ai generated by endpoints in this family.

Veo 3.1 Fast (Text To Video) sample outputInspect →

06 / OPERATIONAL LIMITS

Rate Limits, Concurrency & Upload Constraints

Production enforcement limits enforced by the E2X gateway, worker dispatch, and Cloudflare R2 upload pipeline.

Account API rate tiers
Free: 10 req/min (100/hr, 1,000/day, 2 concurrent) · Pro: 100 req/min (5,000/hr, 50,000/day, 10 concurrent) · Enterprise: 1,000 req/min (50,000/hr, 500,000/day, 50 concurrent)
Default per-model concurrency
5 concurrent jobs per user/model unless overridden per model version (`CONCURRENT_JOBS_EXCEEDED` on overflow)
Gateway global burst protection
10,000 requests/minute gateway cap; fail-closed fallback limiter of 5 req/60s if Redis is unreachable
Batch generations per job
Maximum 16 outputs per POST /v1/jobs/submit request (`GENERATIONS_LIMIT_EXCEEDED` above 16)
Pre-flight estimate & R2 upload limits
POST /v1/jobs/estimate: 60 req/5 min · POST /v1/uploads/presign: 60 req/10 min, max 200 pending uploads, 10 MB images (JPEG/PNG/WEBP/GIF), 100 MB videos (MP4/WEBM)
Webhook delivery policy
HTTPS-only public endpoints (private/loopback SSRF blocked), 10s timeout per attempt, up to 5 attempts (0s, 1s, 4s, 16s, 64s), HMAC-SHA256 `X-E2X-Signature` header

Full per-tier quotas and retry guidance: Rate Limits & Concurrency Documentation →

07 / BENCHMARK COMPARISON

Unit-Normalized Pricing vs. fal.ai

Side-by-side comparison at default unit settings (pricing verified through 24 Aug 2026). Endpoints without a published competitor price render "—".

EndpointCapabilityE2X Default RateSourceCompetitor RateSavingsChecked
google/veo-3.1-fast/start-end-frame-to-videoStart End Frame To Video$0.10/sfal.ai$0.10/s—24 Aug 2026
google/veo-3.1-fast/image-to-videoImage To Video$0.015/sfal.ai$0.15/s90%24 Aug 2026
google/veo-3.1-fast/reference-to-videoReference To Video$0.015/sfal.ai$0.15/s90%24 Aug 2026
google/veo-3.1-fast/text-to-videoText To Video$0.015/sfal.ai$0.15/s90%24 Aug 2026

08 / FAQ

Frequently Asked Questions — Veo 3 API

Verified answers reflecting live catalog models, pricing multipliers, and platform behavior.

Does E2X offer Veo 3, Veo 3.1, or Veo 3.1 Fast?
E2X serves google/veo-3.1-fast across four capabilities: text-to-video, image-to-video, reference-to-video, and start-end-frame-to-video. Non-fast veo-3 and standard veo-3.1 endpoints are not offered in the catalog.
What is the difference between $0.01/s and $0.015/s on Veo 3.1 Fast?
The base second rate is $0.01/s with has_sound set to "false" at 720p or 1080p. Because has_sound defaults to "true" (a 1.5x multiplier), the default rate returned by the catalog is $0.015/s at 720p/1080p and $0.0225/s at 4K with sound.
How is Veo 3.1 Fast start-end-frame-to-video priced?
google/veo-3.1-fast/start-end-frame-to-video uses flat per-second billing at $0.10/s across 4s, 6s, or 8s clips, compared to $0.01/s–$0.0225/s on text-to-video, image-to-video, and reference-to-video.
Which clip durations and aspect ratios does Veo 3.1 Fast accept?
Text, image, reference, and start-end-frame endpoints accept duration values of "4", "6", or "8" seconds in either 16:9 widescreen or 9:16 vertical framing.
How much cheaper is E2X Veo 3.1 Fast than fal.ai?
For text-to-video, image-to-video, and reference-to-video, E2X default pricing ($0.015/s with sound at 720p/1080p) is 90% below the $0.15/s competitor price checked on fal.ai.
How can I check the exact cost before running a Veo 3.1 Fast job?
Send the same model slug and input object to POST /v1/jobs/estimate (rate-limited at 60 requests per 5 minutes) to compute exact costMicroCents prior to job submission.