Create your MCP key.
Generate a Filtrix MCP key in your Filtrix AI account.
Agent-First Generation Stack
Image & Video Generation Built for AI Agents
Generate images and videos with one integration. Built for agent workflows with fast generation modes, reliable retries, task tracking, and unified credits.
- Image + Video in one platform
- Multiple generation modes for speed and quality
- Production-ready reliability for automated workflows
Send this to your agent:
Visit https://app.filtrix.ai/agent-guide.md to get skills.Launch Flow
Go Live in Minutes
Connect your agent client.
Install open-source skills and attach Filtrix MCP through mcp-remote.
Run your first generation task.
Call generation tools and poll task status for reliable automation loops.
Install open-source skills
npx skills add Filtrix-AI/filtrix-skillsCursor / Claude Desktop MCP setup
{
"mcpServers": {
"filtrix": {
"command": "npx",
"args": [
"-y",
"mcp-remote@latest",
"https://mcp.filtrix.ai/mcp",
"--header",
"Authorization:\${AUTH_HEADER}"
],
"env": {
"AUTH_HEADER": "Bearer <YOUR_FILTRIX_MCP_KEY>"
}
}
}
}Multi-Agent Compatibility
Compatible with OpenClaw, Claude Code, Codex, Cursor, and more
One Filtrix skill set runs across multiple agent frameworks, so you can keep one generation workflow and reuse it everywhere.
OpenClaw
Claude Code
Codex
Cursor
What Filtrix Skills Add
- Built-in optimized prompts tuned for image and video generation tasks.
- Reusable workflow templates for automation, reliable retries, and stable runs.
- Batch-ready pipelines for continuous production and rapid iteration.
Use Cases
Built for Real Agent Production Scenarios
Web Design Delivery
No more manual preview images, illustrations, or logos. Your agent can generate and refresh visual assets while building pages.
YouTube & Blog Visuals
Auto-produce thumbnails, article covers, and section visuals so every video and blog post ships with complete media assets.
UGC Marketing at Scale
Use agents with Filtrix skills to batch-produce UGC marketing videos for automated production, testing, and iteration.
Core Capabilities
Everything Your Agent Needs to Create
Text to Image
Turn prompts into production-ready images in seconds.
Text to Video
Generate video tasks from text and track progress end to end.
Task Status Tracking
Check generation state programmatically and build stable pipelines.
Unified Credits
One balance across image and video workflows.
Generation Modes
Choose the Right Mode for Every Task
gpt-image-1
Best overall quality for polished outputs.
nano-banana
Fast generation for iteration-heavy workflows.
nano-banana-2
Advanced controls for higher-quality creative tasks.
FAQ
Frequently Asked Questions
How is billing calculated?
Usage is charged per generation task based on selected mode. Image and video use one shared credit balance so agents can route work without juggling wallets.
How do retries avoid duplicate charges?
Use idempotent task handling in your workflow and query task status before retrying. This keeps agent loops stable and prevents accidental duplicate generations.
Which mode should I choose first?
Start with nano-banana for fast iteration and switch to gpt-image-1 or nano-banana-2 when you need polished output or tighter creative control.
How do I track long-running video tasks?
Submit a video generation request, store the task id, then poll status in your pipeline until completion to safely coordinate downstream agent steps.