Generate hyper-realistic avatar and UGC talking-head videos. Create up to 60 seconds of production-ready video by simply selecting an avatar and writing the script.
Sume achieves production-ready video generation through dynamic multi-model orchestration.
Behind a single reliable model and API surface, Sume coordinates specialized generation steps for model routing, visual reasoning, reference manipulation, and avatar control.
Use Sume directly, or connect it to agents through the same reliable generation layer.
Use it by your own
Integrate Sume directly through the Developer API. Submit structured avatar video jobs, pass scripts and references, track status, and retrieve production-ready media from your own product.
API quickstart(opens in a new window)
Use it with your agent
Connect Sume to agent workflows through the CLI. Agents can call the same generation layer, coordinate creative inputs, and use completed media in downstream tasks.
CLI quickstart(opens in a new window)Use the docs to build through the Developer API and CLI. Use it directly, or let your agent use it.
Set up Sume, understand the core workflow, and run your first developer integration against the Sume platform.
Build with durable jobs, model runs, polling or webhooks, first-party media URLs, assets, results, errors, and recipes.
Learn how authentication, workspace-scoped API keys, catalog discovery, usage tracking, and dashboard observability fit together.
Use the avatar and avatar-video model docs to create controllable UGC talking-head generation workflows.
Install, authenticate, discover the catalog, manage jobs and assets, run generation workflows, and use --json or --agent flags.
Track usage, understand credit spend, inspect job costs, and manage billing controls for production integrations.
Book a 15-minute interview with Sume's co-founder. You'll receive $50 in free credits, and the session is simply a casual conversation about what you're building and how you'd like to use Sume. For enterprise inquiries, contact david@sume.com.
