Jimeng First Frame to Video

Problem

When you already have a first-frame image and want to extend it into a dynamic video, calling the Jimeng image-to-video API directly can involve managing API keys, parameter uploads, async task polling, and error handling. This skill collapses that workflow into a single dlazy jimeng-i2v-first command, which is useful for scripts, automation pipelines, and agent flows that need consistent output.

How It Works

• Input: Provide a first-frame image and text prompt. Local files are uploaded to dLazy media storage so the backend model can read them.
• Invocation: The skill acts as a thin client for the dLazy hosted API, sending the prompt and parameters to api.dlazy.com for inference.
• Output: Synchronous calls return results; async mode returns a task.generateId field for continued status polling.
• Error handling: It distinguishes missing API key, missing parameters, local file read failures, insufficient balance, server errors, and content-policy blocks, which helps automation branches handle each case.

Boundaries

This skill wraps a remote generation service rather than running local inference or offline processing. It involves cloud file uploads and hosted output URLs, so service terms, balance, and API key permissions matter. For sensitive assets, evaluate upload compliance first and verify that the generated output meets content policies.