MuAPI

@MuAPI list all image-generation models in your catalog that support 1024×1024 output and show me their per-image pricing.

The catalog includes 14 models supporting 1024×1024 generation. DALL·E 3 Standard costs $0.040 per image; Stable Diffusion XL costs $0.008; Midjourney v6 costs $0.025; Flux Pro costs $0.055. Each entry shows the model family, provider, and whether it supports batch requests. This is a representative sample—actual catalog size and pricing depend on your MuAPI subscription tier and the provider's current rate card.

This example demonstrates catalog browsing, which is read-only and free. Before running expensive generation jobs, always call the estimate endpoint with your exact parameters—dynamic pricing models (like some Anthropic or OpenAI variants) can vary by token count, and this list shows base rates only. No API key scope beyond read access is required.

@MuAPI estimate the cost to generate 50 images using Stable Diffusion XL with the prompt 'minimalist product photography, white background' at 1024×1024 resolution.

Estimated total: $0.40 USD (50 images × $0.008 per image). This estimate assumes default sampling steps and no upscaling. If you enable the refiner pass or request higher step counts, the per-image cost will increase—re-run the estimate with those parameters before committing. Estimation calls are free and do not consume account credits.

Cost estimation is the safety check before batch generation. It surfaces the exact charge for your parameters, which matters when a model has tiered pricing or optional add-ons. Always estimate first if you're unsure—MuAPI bills immediately on generation, and there's no undo for a misconfigured 500-image run.

@MuAPI generate 3 variations of 'cozy coffee shop interior, warm lighting, editorial style' using DALL·E 3 and save the URLs so I can review them.

Generated 3 images via DALL·E 3 Standard (1024×1024). Total charge: $0.12 USD, deducted from your MuAPI account balance. Image URLs: [url1], [url2], [url3]. Each URL expires in 60 minutes per OpenAI's policy—download them now if you need permanent copies. The AI can compare the outputs and suggest which composition best matches your brief, but it cannot see the images directly without you sharing them.

This is a billable write operation that consumes account credits immediately. The MCP returns temporary URLs, not stored assets—if you need the images later, download them within the expiration window. Requires an API key with generation permissions and sufficient account balance. Rate limits apply per your MuAPI plan tier.

When you're building an AI feature and need to compare model economics

A 3-person startup is adding AI image generation to their SaaS product and needs to pick a model that balances quality and unit cost. This MCP wins here because the Estimate Model Cost tool lets you price a generation before you run it—critical when you're testing 8 different models and don't want to burn $40 in API credits just to find out which one fits your margin. The List Catalog Models tool shows pricing across the full native catalog, not just the OpenAI-compatible subset, so you see options your competitors might miss. The threshold: if you're only using one model in production and already know its pricing, this MCP is overkill. But if you're in the discovery phase or running A/B tests across providers, the cost-estimation loop pays for itself in the first afternoon.

When support needs to generate images without touching production keys

A 6-person e-commerce team gets 15 customer requests a week for custom product mockups, and the support reps currently Slack the design lead every time. This MCP is borderline. The Generate Image tool does let support run generations directly from the AI workspace without exposing production API keys, which cuts the median response time from 4 hours to 10 minutes. But MuAPI's catalog is optimized for developers comparing models, not for support reps who just need one reliable image endpoint. If your support volume is under 20 requests a week, the juice isn't worth the squeeze—just give the design lead a Slack command. Above that threshold, or if you're fielding requests across multiple image styles, the MCP starts to make sense because the List Models tool surfaces the right model for each request type.

When you're scoping AI work and need live pricing in the sales cycle

A 10-person creative agency is pitching a campaign that includes AI-generated visuals, and the client wants a cost breakdown before signing. This MCP wins because the Get Catalog Model tool pulls live pricing and schema details during the pitch meeting—no waiting for the dev team to run test calls or dig through API docs. The agency can show the client exactly what a 500-image batch will cost across three model tiers, then adjust the scope in real time if the budget tightens. The trade-off: if your agency only pitches AI work once a quarter, you don't need an MCP for this—just bookmark the pricing page. But if you're quoting AI features weekly and your win rate depends on fast, accurate estimates, this MCP turns pricing research from a 2-day dev task into a 5-minute live demo.