venice

name	venice
description	Use this skill when the user wants to query a specific AI model, generate media (video/image/audio), or ask uncensored questions. Trigger phrases include "ask deepseek", "ask qwen", "ask llama", "generate video", "generate image", "create audio", "uncensored", or model names.

Venice.ai Tools

Execute these commands directly when the user requests model queries, media generation, or uncensored queries.

Query Specific Models

Use venice-ask -m <model> to send a question to a specific Venice model:

venice-ask -m deepseek-r1-671b "your question"
venice-ask -m qwen-2.5-coder-32b "your question"
venice-ask -m llama-3.3-70b "your question"
venice-ask "your question"  # Uses venice-uncensored by default

Available models:

Model ID	Best For
deepseek-r1-671b	Deep reasoning, math
qwen-2.5-coder-32b	Code generation
qwen3-235b	Large context
llama-3.3-70b	General purpose
venice-uncensored	Unrestricted queries (default)

Action required: When user says "ask deepseek about X" or "use qwen to answer Y", run:

venice-ask -m deepseek-r1-671b "X"

Media Generation

Video (async - returns job ID):

venice-generate video "description of video"
venice-generate status <job-id>  # Check progress and download

Image (instant):

venice-generate image "description of image"

Audio/Speech:

venice-generate audio "text to speak" --voice af_sky
venice-generate voices  # List available voices

Output files are saved to ~/venice-output/

Instructions

1. Execute commands directly - Don't just show commands, run them
2. Show output - Display command results to the user
3. Video is async - Tell user the job ID and explain they can check status later