wtf is venice-media-skill?

Venice Media Skill is a Python bridge that lets an existing AI command line tool generate images, videos, speech, and music through the Venice API, without replacing that tool. Instead of becoming its own chat agent, it works underneath assistants like Kimi Code, Codex, Claude Code, Gemini CLI, or OpenCode, letting them call out to Venice for media generation while the original assistant keeps handling the conversation. Once installed, the tool covers a wide set of media tasks: generating, editing, and upscaling images, removing image backgrounds, generating and retrieving videos including edited or extended clips, converting text to speech, generating music and other audio, and transcribing spoken audio into text. It can also look up which AI models are currently available directly from Venice, so it does not rely on a fixed, potentially outdated list. The project pays close attention to safety and cost control around API usage. Your Venice API key is only ever read from an environment variable and is never written into any saved file. If a video or audio job times out, it returns a resumable job ID instead of quietly resubmitting the request and charging you twice. For paid video and audio generation, the tool can return a price quote first and wait for approval before actually running the job. Generated files also come with metadata files describing what was made, unless a manifest specifically turns that off. Installing it requires Python 3.11 or newer and a Venice API key, which you generate from Venice's own settings page. You can install it by cloning the repository and using pip, or by running a provided install script for macOS, Linux, Windows Subsystem for Linux, or Windows PowerShell, which sets up an isolated environment and a command line launcher automatically. A built in doctor command checks whether your local setup and your connection to the Venice API are both working correctly. The project is released under the MIT License.