Blog
AI Image & Video API Guides
Hands-on model tests, production tutorials, prompt techniques, and current API pricing for generative media teams.
Seedance 2.0 Mini Prompt Recipes: Copy-Paste Prompts With Real Outputs
Five copy-paste seedance-2.0-mini prompt recipes, each paired with the real clip it generated on the live API (no borrowed demos). Covers text-to-video and image-to-video (first-frame control), plus camera moves, visual style, fast action, and native synced audio. Verified model facts: 4–15s clips, 480p/720p, duration + resolution required, aspect_ratio optional, T2V and I2V on one async task endpoint. Pricing is per second: $0.068/s at 480p, $0.147/s at 720p; a reference video lowers the rate. Includes a ready-to-run Python snippet (submit, poll, download) you can paste and use today.
Seedance 2.5 vs Seedance 2.0: What's New, Release Date, and Pricing
Seedance 2.5 is the next-generation video model previewed at the Volcano Engine FORCE 2026 conference (June 23, 2026). These are announced specs, not tested results. Headline upgrades over 2.0: up to 30s one-shot generation, native 4K (10-bit), up to 50 multimodal references, 3D blocking for camera control, and local editing with native audio. Expected to open for API access in early July. As of June 29, 2026, it is not yet publicly available, and no official price has been published. What you can use right now: Seedance 2.0 on hiapi (4-15s clips, up to 1080p, image-to-video, synced audio) on the same async task API. To be first to integrate: build your pipeline on Seedance 2.0 today, then swap the model name on launch day. One API key covers both.
Generating Ecommerce Product Images with qwen-image-2.0 via the hiapi API
qwen-image-2.0 on hiapi produces ecommerce-grade images with clean Chinese-text rendering for $0.025 per image — the only model on the platform that puts both Chinese headlines and English subtitles in the same render reliably. Input schema is unique to this model: field is 'size', separator is '*', and only five values are accepted ('2688*1536', '2368*1728', '2048*2048', '1728*2368', '1536*2688'). Other models on hiapi accept 'aspect_ratio' — qwen-image-2.0 does not. A 200-SKU cross-border drop (hero + lifestyle + variant grid + banners + listing cards) totals around 810 renders at ~$20 — the cost band that makes 'generate everything' cheaper than commissioning a partial photoshoot.
Free Text-to-Video AI API: Realistic Free Options and How hiapi Compares
There is no provider in 2026 that gives you production-grade text-to-video at zero marginal cost with no friction — only three real routes that get close: Hugging Face's free Inference Provider credits, self-hosting an open model, or a small per-second paid API. Hugging Face's free tier is the closest thing to a true free API: routed OpenAI-compatible access to open video models on Replicate, fal, and Novita, with monthly credits good for roughly 5–50 short clips per month. Self-hosting Wan 2.1/2.2, HunyuanVideo, CogVideoX, or LTX-Video on owned or rented GPUs is genuinely cheaper per clip at volume — but only after you absorb the model selection, quantization, and ops cost. For 20+ five-second 720p clips per month, hiapi's wan2.7-video/text-to-video@pro at $0.10/sec ($0.50/clip, $40/mo for 80 clips) usually beats both routes once you price in your engineering time.
Image Generation API Rate Limits Explained: What to Expect and How HiAPI Handles Bursts
Image APIs gate traffic on three axes — RPM, RPD, and concurrency — and concurrency is almost always the real bottleneck for image generation. A 429 is not a 401: read the error body, grab the request_id, and only back off on actual rate-limit codes. Cap client-side concurrency with a semaphore before adding retries; exponential backoff with full jitter handles transient bursts without thundering-herd. HiAPI's POST /v1/tasks split makes rate-limit math simpler: cheap create calls + provider-side queueing absorb bursts behind one Bearer key. Honor Retry-After when present; separate keys per environment so staging traffic doesn't eat production's concurrency budget.
hiapi output URL expired: download before it expires
hiapi `/v1/tasks` returns a signed `output[*].url` with a short `expireAt` — minutes, not days. Once `expireAt` passes the URL 403s; the task record stays but the file is gone from that link. Download the bytes inside the same polling loop and upload them to your own storage (R2/S3/CDN). Never store the signed URL in your DB or hot-link it into customer-facing pages. Re-running the task to recover a lost asset is a new billable job — cheaper to download once.
![Generating Ecommerce Product Images with FLUX.1 [schnell] via the hiapi API](/_next/image?url=https%3A%2F%2Fstatic.hiapi.ai%2Fblog%2F_default%2Flight-cover-12.png&w=3840&q=75)
Generating Ecommerce Product Images with FLUX.1 [schnell] via the hiapi API
flux-schnell on hiapi produces ecommerce-grade product images (packshot, lifestyle, top-down, variants) for $0.005 per image and about 30 seconds end-to-end — the right speed/cost band for catalog work. schnell's input schema rejects 'resolution' — send only 'aspect_ratio'. Other image models on hiapi accept both. For a 200-SKU catalog at 4 shots each (800 images) total cost is $4. Premium models for the same job land at $40–$136.
Happyhorse Image to Video API: Working curl and Python Examples (hiapi)
Animate one still image into a 720p or 1080p MP4 with `happyhorse-1.1/image-to-video` on the hiapi tasks endpoint. Working curl and Python examples that pass schema validation today, with the actual field names and value ranges the API enforces. Production callback pattern (`callback.url` + `when: "final"`) plus the exact error strings you will hit, so you can match on them in your handler. Guidance on downloading the output URL before its `expireAt` timestamp drops it from temp storage.
When a hiapi /v1/tasks Job Hangs or Times Out: Diagnosis and Retry
Tasks that sit in `pending` or hit `timeout` almost always trace back to four causes: queue backlog, an underlying model failure, oversized inputs, or an unreachable callback URL. Diagnose by polling `GET /v1/tasks/{id}` and reading `status` + `error.code` before you retry anything. Retry with exponential backoff (e.g. 5s → 15s → 45s, max 3 tries) and always create a **new** task — never POST the same idempotency body twice without a key. If a task stays in `pending` for over 90 seconds with no `error`, it's almost always a queue backlog — back off, don't hammer. For callback delivery issues, verify the URL is publicly reachable and returns 2xx within 10s; the hiapi platform does not retry indefinitely.
HappyHorse 1.1 Text-to-Video API via HiAPI: a working curl + Python tutorial
Call Alibaba's HappyHorse 1.1 text-to-video model through HiAPI's unified /v1/tasks endpoint with a working curl request, a Python polling client, and a production-ready callback pattern. The exact input schema (prompt, resolution, aspect_ratio, duration) plus the two non-obvious 400s — audio and seed are not accepted — are spelled out so the first request actually goes through.
Build with the models you just read about.
Compare image, video, and audio models, then use the same API workflow in production.