ZSky Max API Is Live: The Best AI Video + Image API for Creators in 2026
The ZSky Max API ships today. Every Max-plan subscriber now has programmatic access to the same generation engine that powers zsky.ai/create — video with synchronized audio, image, and the asynchronous job pipeline — through a clean REST API.
The headline numbers: 300 video generations and 1,000 image generations per month included in the base $99 plan. Opt-in overage at $0.50 per video and $0.05 per image, capped by a spend ceiling you set yourself between $5 and $1,000. Wallet balance rolls over forever — you never lose what you have paid for.
This guide walks through what's in the API, why we think it is the best option for creators and small teams in 2026, and how to make your first call in under five minutes.
What's in the Max API
The API exposes the production generation engine through three groups of endpoints:
POST /v1/generate— submit a text prompt and a generation type (videoorimage). Returns a job ID.GET /v1/job_status/{id}— poll job status. Returns the result URL on completion.GET /v1/me/overage,POST /v1/me/overage/toggle,POST /v1/me/overage/cap— read or update your overage settings, spend cap, and wallet balance.
Auth is a single header: X-API-Key: zs_live_…. You issue and rotate the key from /settings in the "API & Overage" section that appears once you're on the Max plan.
Output: video at 1080p with synchronized audio, image at 1080p. No watermark on Max-plan output — the free-tier wordmark applies only to free-tier web users. API output is delivered clean for commercial use.
6 reasons ZSky Max API wins in 2026
Most video APIs return silent clips. You then bolt on a second API call to a text-to-speech service, then a third call to align, score, and mux. ZSky returns the video with audio baked in — one prompt, one job, one URL.
Most APIs sell a pre-paid balance that gets burned down per second of compute. The Max plan gives you a flat monthly base of 300 video generations and 1,000 image generations. No balance math, no surprise drain when a longer prompt costs more.
Opt in to overage and your unused balance carries month to month — no expiry, no decay. Cancel the Max plan and the existing wallet stays spendable until depleted. Customers shouldn't lose what they paid for.
Pick a monthly cap between $5 and $1,000. The API enforces it on every call. A bug in your retry loop can never produce a five-figure surprise bill — once the cap hits, calls return 429 and stop costing.
Some providers run a stripped-down "API model" alongside their consumer product. ZSky doesn't. The Max API hits the same GPUs and the same pipeline as zsky.ai/create. What you ship in your app matches what users see on the web.
Every wallet deduction, replenish, refund, alert, and cap change writes a row to your audit ledger. You can reconcile every charge against a generation. End-of-month billing is itemized, not a black box.
The roadmap layers more on top: programmatic access to the AI Creative Director (currently a web-only conversational creative partner on Pro+), image-to-video and image-to-image endpoints, webhook callbacks instead of polling, and SDKs for Python, JavaScript, and Go.
Quick start
Five minutes from key to first video.
1. Get your key
- Subscribe to Max at zsky.ai/pricing ($99/mo monthly or $79/mo billed annually).
- Visit zsky.ai/settings → "API & Overage".
- Click Rotate Key to issue your first key. Copy it into your secret store.
2. Submit a generation (Python)
import os, time, requests
KEY = os.environ["ZSKY_API_KEY"]
H = {"X-API-Key": KEY, "Content-Type": "application/json"}
# Submit
r = requests.post(
"https://zsky.ai/v1/generate",
headers=H,
json={
"type": "video",
"prompt": "A neon-lit night market in Tokyo, slow dolly forward, ambient crowd sound",
},
).json()
job_id = r["job_id"]
# Poll
while True:
s = requests.get(f"https://zsky.ai/v1/job_status/{job_id}", headers=H).json()
if s["status"] == "completed":
print("video:", s["result_url"])
break
if s["status"] == "failed":
raise RuntimeError(s.get("error"))
time.sleep(2)
3. Or with curl
curl -X POST https://zsky.ai/v1/generate \
-H "X-API-Key: $ZSKY_API_KEY" \
-H "Content-Type: application/json" \
-d '{"type":"image","prompt":"A minimalist espresso bar in Kyoto, golden hour"}'Video jobs typically complete in ~30 seconds. Image jobs typically complete in ~2 seconds.
How it compares to Sora, Runway, Pika, and Luma APIs
The closest competitive set on the video side is Sora, Runway, Pika, and Luma Dream Machine. Most charge per second of compute or sell time-bound pre-paid packs. Most return silent video. None bundle a self-set spend cap with a rollover wallet.
| Provider | Video w/ synced audio | Pricing model | Spend cap | Wallet rollover |
|---|---|---|---|---|
| ZSky Max API | Yes — one call | $99/mo + opt-in overage | You set $5–$1,000 | Forever |
| Sora (via OpenAI) | Limited, varies | Plan-tier + metered usage | Plan-tier ceiling | Resets each cycle |
| Runway Gen-4 / Alpha | No (silent video) | Per-second metered usage | Pre-paid balance | Balance expires |
| Pika | No (silent video) | Pre-paid packs | Pre-paid balance | Balance expires |
| Luma Dream Machine | No (silent video) | Per-generation metering | Pre-paid balance | Balance expires |
| Replicate (image models) | N/A (image only) | Per-second compute | Account balance | Balance carries |
| OpenAI Images API | N/A (image only) | Per-image flat rate | Account balance | Balance carries |
Two practical effects: (1) you can build a creator-app feature where a user types one prompt and gets a finished video-with-audio asset back, without writing TTS-mux code; (2) you can ship that feature with a hard cost ceiling so a viral moment doesn't put a five-figure bill on your card.
Pricing breakdown
| Item | Cost |
|---|---|
| Max plan — monthly | $99 per month |
| Max plan — annual | $79 per month, billed annually |
| Included video generations | 300 per month |
| Included image generations | 1,000 per month |
| Overage — video | $0.50 per generation (opt-in) |
| Overage — image | $0.05 per generation (opt-in) |
| Spend cap | $5 to $1,000 / month (you set it) |
| Wallet rollover | Forever — unused balance never expires |
| Watermark on output | None on Max-plan output |
| Commercial rights | Full commercial use of generated assets |
Compared with a per-second-metered provider charging roughly $0.05–$0.10 per second of 720p–1080p video, a 6-second clip costs $0.30–$0.60 each — and the balance typically expires. The Max plan's flat $0.50 per video stays predictable regardless of length within our supported range, with no expiry on the rollover wallet.
Start building with Max API
Subscribe to Max, issue your key, and ship your first video-with-audio generation in five minutes.
See Max pricing →Frequently asked questions
What is the ZSky Max API?
The ZSky Max API is a self-serve generation API for paying customers on the Max plan ($99/month or $79/month billed annually). It includes 300 video generations and 1,000 image generations per month at no extra cost, with an optional pay-as-you-go overage at $0.50 per video and $0.05 per image. Video output includes synchronized audio in a single call.
How much does ZSky Max API cost?
The Max plan is $99 per month billed monthly, or $79 per month billed annually. Your base allowance is 300 videos plus 1,000 images per month. If you enable opt-in overage, additional generations cost $0.50 per video and $0.05 per image, capped by a monthly spend limit you set between $5 and $1,000. The overage wallet rolls over forever — you never lose what you have paid for.
How is the ZSky Max API better than Sora, Runway, Pika, or Luma APIs?
Four ways. First, video output includes synchronized audio in one call — no separate text-to-speech or scoring step. Second, you get 300 video generations per month included in the base plan, not metered from a pre-paid balance. Third, the optional overage wallet rolls over month to month and never expires. Fourth, you set your own spending cap between $5 and $1,000, so a bug in your code can never produce a surprise five-figure bill.
Does ZSky Max API output have a watermark?
No. Max plan output, including video and image generations produced through the API, is watermark-free. The free-tier wordmark applies only to free-tier web users. Paid-tier output across web, mobile, and API is delivered clean.
What endpoints does the ZSky Max API expose?
The core endpoint is POST /v1/generate, which accepts a text prompt and a generation type (video or image) and returns a job ID. You poll GET /v1/job_status/{id} to check progress and retrieve the result URL. The /v1/me/overage endpoints let you read or update your overage toggle, cap, and wallet balance. Authentication uses an X-API-Key header issued from your settings page.
What happens if a generation fails?
Internal failures (server error, GPU timeout) are automatically refunded — you are not charged. Safety blocks, where the request is rejected for policy reasons, do consume the wallet reservation because the GPU work was performed. Every wallet motion writes a row to your audit ledger, viewable in your settings page, so you can reconcile usage and charges precisely.
Can I use ZSky Max API output commercially?
Yes. Max plan customers receive full commercial usage rights to API-generated video and image output. You can use it in products, marketing, client work, merchandise, and editorial. Always confirm the specific terms in your subscription agreement.
How do I get a ZSky Max API key?
Subscribe to the Max plan at zsky.ai/pricing, then visit zsky.ai/settings. The "API & Overage" section appears for active Max subscribers. Click "Rotate Key" to issue your first key, then copy it into your application. You can rotate the key at any time if it leaks.
Ship your first video-with-audio in five minutes
Subscribe to Max, issue your key from settings, and run the Python snippet above.
Get Max →