Official ByteDance API partner

Give your AI agent
a face it never forgets.

Stop rebuilding your character for every scene. One API handles it all.

500 free credits No credit card Cancel any job · full refund
claude-code · synts mcp
connected
> make a tokyo drift edit with @character_1 and @character_2 with seedance 2.0
Using synts mcp
synts.generate · seedance-2-0
refs: @character_1 @character_2 → 2 uploadIds
job_7f3ebcb6 · 30s clip · ~2min
▶ 0:30 · 1080p same faces as @character_1 + @character_2
synts.ai
×
BytePlus
Official partner of BytePlus, a ByteDance company. The AI behind TikTok, CapCut, Doubao.
100%
Refunds on all failed generations
Unlimited
Character variables
MCP
Claude Code + Cursor ready
1 key
Every ByteDance model

MODELS

All outputs generated via Synts API. No post-processing.

Primitives

Character memory. Namespace control.
ByteDance doesn't build this. We do.

Raw model access is table stakes. Variables, jobs, refunds, chaining. Everything you'd otherwise stitch together yourself.

Variables

Save any asset by slug. Reference @character1 in any prompt, any session, any model. Same face, guaranteed.

Survives sessions
Works in every model
Unlimited refs per account
Agent-safe (stable IDs)

Unified endpoint

Every ByteDance model through POST /api/generate. One credit balance. One auth header. One SDK.

vs stitching 5+ vendor APIs

Multi-step pipelines

Every output returns a reusable upload_id. Feed Seedream into Seedance into OmniHuman. Zero re-uploading. Zero URL handling.

Agents can't hallucinate IDs that don't exist

Job management

Full PENDING → PROCESSING → COMPLETED state machine. Filter, cancel, retry, webhook.

Cancel = 100% credit refund, auto

Presets

Bundle prompt + model + resolution + references into one preset_id. Reproducible outputs on every deploy.

Version-control your generations

Cost predictability

Check exact credit cost before every call: GET /billing/credits/costs. No surprise bills at any scale.

1 credit = $0.001 · always

Storage

Upload any asset once and get a CDN-backed uploadId ready for every model call. No S3 plumbing to build yourself.

GET /api/storage/uploads

History

Every generation saved automatically. Filter by model, re-download, replay a prompt, or cancel any in-flight job.

GET /api/generation/history

For coding agents

Ships with TypeScript, Python, and MCP.

Stable upload_ids on every output. Your agents can't hallucinate URLs that don't exist. Variables persist across sessions via @mentions. Zero-config MCP server for Claude Code, Cursor, Windsurf.

Read the docs
Claude Code Cursor Windsurf OpenAI compatible

Why Synts

Other AI platforms are slow,
don't refund, and don't have an API.

We benchmarked the major AI video and image platforms against the four things developers actually care about. The gap is not subtle.

Image gen
10–30s
vs 30s–5min elsewhere
Video gen
30s–3min
vs 4–40min elsewhere
Refunds
Automatic
others: 7-day window, if zero credits used
Real API
Every plan
others: gated, hidden, or none
Capability
synts.ai
API + Agent layer
Other AI platforms
Industry pattern
Image generation speed 10–30 seconds 30 seconds to 5 minutes
Video generation speed 30 seconds to 3 minutes 4 to 40+ minutes
Automatic refund on failed jobs
credits return the moment a job errors
Refund on disliked output
flag a generation, get credits back
Real REST API + SDKs
on every paid plan
TS · Python · curl
gated, hidden, or none
Official ByteDance partner
the inference behind TikTok and CapCut
Persistent character variables
slug-based, work in every model and session
Starts at $19/mo $7–$25/mo
Speed

"Why is my generation stuck at 99%?"

Generation queues balloon during peak hours. Paid plans get throttled after launch novelty wears off. A single video can sit pending for 15, 40, or sometimes a full day before timing out.

Failed jobs

Credits gone, video never delivered

Failed generations consume credits anyway. Some platforms admit the video "did not generate" but cite "compute costs" to deny refunds. One failed test video can burn the majority of a monthly allowance.

Trust · Synts

Official ByteDance partner

Synts runs on BytePlus, the same inference layer that powers TikTok, CapCut, and Doubao. That means enterprise-grade speed, uptime, and model quality, with a direct partnership behind every call. No wrapping, no aggregating, no middleman.

API access

No API, or API only at the top tier

Most consumer-facing AI video tools have no public API at all. Some gate it to the highest plan ($60+/month), some require enterprise sales conversations, and others route through third parties. Building agents on these stacks is brittle.

Synts gives every credit back. Every time.

Failed generation? Auto-refunded. Don't like the output? One-click refund. The same ByteDance models others wrap, with infrastructure that actually finishes the job.

Community · Free forever

Learn Synts and AI,
in public, for free.

A free Skool community where we teach you how to use Synts, every ByteDance model, and the broader AI stack you need to ship. Tutorials, live calls, and a feed where you post work and get answers.

Join the community Free · No card · Open to all
Synts, end-to-end
Agent setup, character variables, multi-step pipelines. From first key to first generation in under an hour.
AI fundamentals
Prompting, tools, agents, video pipelines. The same stack we use to ship Synts itself.
Live builds & Q&A
Real projects in real time. Bring yours, get unstuck, watch how others ship.
An active feed
Post your work, get feedback, share wins. Founders, devs, and creators in one room.
Support

Need a hand? Two ways to reach us.

Real humans. Fast replies. No bots in the loop.

Pay for what you generate. Nothing else.

One credit. One balance. Every model. No hidden costs.

$0 Free tier
500 free credits to start
No card required. Start in 60 seconds.
Get started

Starter

$19 first month
Renews at $25/mo
15,000 base credits/mo
+ 2,500 reward credits
Get Started
  • 15,000 base credits
  • 2,500 reward credits
  • Up to 2K images / 720p video
  • 10 GB storage
  • 30-day history
  • API access (5 keys)
  • Analytics dashboard
MOST POPULAR

Pro

$35 first month
Renews at $49/mo
45,000 base credits/mo
+ 12,500 reward credits
Go Pro
  • 45,000 base credits
  • 12,500 reward credits
  • Up to 4K images / 1080p video
  • 100 GB storage
  • 90-day history
  • Priority support
  • Custom branding
BEST VALUE

Max

$65 first month
Renews at $129/mo
75,000 base credits/mo
+ 25,000 reward credits
Go Max
  • 75,000 base credits
  • 25,000 reward credits
  • Up to 4K images / 1080p video
  • 250 GB storage
  • 180-day history
  • Priority support
  • All Pro features
FAQ

Common questions

On paid plans, full ownership and commercial rights transfer to you. Marketing, products, client work, resale, anything. When you create content on Synts, you're the one making the creative decisions: choosing the prompt, selecting parameters, and directing the output. We treat you as the owner. Free tier content is licensed for personal use only.

Yes. You are solely responsible for the content you create using Synts and how you use it. Synts provides AI tools and you provide the creative direction. That means it's on you to make sure your content doesn't infringe anyone's copyright, defame a real person, violate data protection laws, or break any other applicable laws. We implement safety filters and content moderation, but no automated system is perfect. If your use of the platform results in a legal claim, you (not Synts) bear that responsibility. See our Terms of Service for full details.

Most tools give you one model or a raw API with no structure around it. Synts wraps the five ByteDance models — Seedream, Seedance, OmniHuman, DreamActor, Doubao TTS — with a workflow layer: character memory, job management, presets, and automatic refunds. You get one credit balance, one API key, and outputs that chain into each other without any glue code. The models are ByteDance's. Everything around them is ours.

Yes — the whole product is an API. Starter gets 5 keys at 60 RPM, Pro gets 20 keys at 300 RPM, Business scales to unlimited keys at 3,000 RPM. REST, with TypeScript and Python SDKs, plus a zero-config MCP server for Claude Code, Cursor, and Windsurf. Docs are in your dashboard the moment you sign up.

Cancel anytime from your account settings. No lock-in, no questions asked. If you subscribe and haven't used any credits within the first 24 hours, you can request a full refund at support@synts.ai. Once credits are used, the subscription is non-refundable. It's your responsibility to manage your subscription. We don't issue refunds for forgetting to cancel.

Your prompts and generated content are yours. We don't use your prompts to train models, and we don't share your content with third parties. Generated images and videos are stored securely and only accessible to your account. You can delete your data at any time.

Any paid subscriber can buy top-up credit packs from their dashboard: Lite ($10 / 6,500 credits), Standard ($25 / 18,000 credits), or Plus ($50 / 40,000 credits). Business users have larger packs available. Pack credits never expire. Monthly subscription credits reset each billing cycle.

1 credit = $0.001, so 1,000 credits = $1. Every action on the platform costs a flat number of credits: image generation starts at 50 credits, video generation at 50 credits per second, and costs scale with resolution. Your monthly plan refreshes credits each billing cycle. You can also top up with one-time credit packs that never expire.

Log in daily, earn free credits. Higher plans earn more per day. Starter earns up to 5,000/mo, Pro up to 15,000/mo, and Max up to 25,000/mo. Hit login streaks at 3, 7, 14, and 30 days for extra bonuses. Reward credits reset each billing month, so use them while you have them.

Subscription credits (both your base credits and reward credits) reset every billing cycle. Unused ones don't roll over. Credit packs purchased separately never expire and stay in your account until you use them.

Images up to 4K, video up to 1080p, lip sync from any audio or text input, and characters that stay consistent across every generation. All five ByteDance models, one credit balance, one API key. No switching between tools.

Get your API key from settings, drop the docs URL into your agent's context, and you're done. Your agent can read every endpoint, call generations, manage jobs, and reference stored characters — all without you writing a single integration layer.

Email support@synts.ai or hit us on Discord — link in the footer. Business and Enterprise plans get priority response and a dedicated support channel. We're a small team and we actually respond.

Synts × ByteDance

One API key.
Your agent never forgets a face.

500 free credits. No card. Key in under 60 seconds.