AG Video Agent Builder
Agno SDK + OpenRouter (google/gemini-3-flash-preview) + Fal mapping
ag.winds-os.com
1) Agent Builder
2) Generated Agents List
3) Video Generations from Agents
Video Instruction
Video Type
Tutorial video (AI video generation tutorial)
Sample video (actual sample/reference output)
Action
Add Instruction to Generate
Pipeline: production brief → build process/models → Fal MCP mapping → generate video → configure app
This screen builds a NEW agent/video flow from your instruction.
Video Generations from Agents
Refresh Video Generations
Video Generation Steps
Generated Agents List
Refresh Agents List
Agent + Selected Run Steps
Run Selected Agent (New Video)
Video Instruction
sent to agent
Video Type
sent to agent
tutorial
sample
Action
Add Instruction to Run Agent
Edit Agent Models / Prompts / Workflow (Selected Agent Template)
Context Tags (Video Generation)
Sent to AI models
sent to agent
Video instruction + video type
Workflow models + options + instructions
Blueprint prompt + scene prompts + narration text + voice/media URLs for Fal steps
Not sent to AI models
local only
Run ID, created time, UI status badges
History tables and editor chrome
Any field marked as "new-agent setup only" during Run Agent flow
Models
Text Model ID
sent to agent
Fal Image Model
sent to agent
Fal I2V Model
sent to agent
Fal LipSync Model
sent to agent
Fal TTS Model
sent to agent
Eleven Voice ID
sent to agent
Options
Scene Count
sent to agent
Scene Duration (s)
sent to agent
Aspect Ratio
sent to agent
I2V Resolution
sent to agent
(model default)
360p
480p
540p
720p
1080p
1440p
4K
Default Video Type
sent to agent
tutorial
sample
Prompts
Instruction Brief Prompt
sent to agent
Blueprint Prompt
sent to agent
App Config Prompt
new-agent setup only
Instructions (one per line)
Instruction Brief Instructions
sent to agent
Blueprint Instructions
sent to agent
App Config Instructions
new-agent setup only
Save Agent Workflow Settings