ImageRouter
Generate images, videos, and text with ImageRouter, inspect credits, validate normal API keys, and discover models. Management-key endpoints are excluded.
Verdict
Common use cases
- Prototype UI mockups with multiple image models
- Generate marketing visuals without vendor lock-in
- Compare video outputs across different generators
- Run complex reasoning tasks on premium LLMs
- Inpaint product photos for A/B tests
Integration
- Vendor
- ImageRouter
- Category
- other
- Auth
- API_KEY
- Tools
- 8
- Composio slug
imagerouter
Tools
- Create Chat Completion
Create a non-streaming OpenAI-compatible chat completion with a paid text model. Call IMAGEROUTER_LIST_MODELS, IMAGEROUTER_GET_MODEL, and IMAGEROUTER_GET_CREDITS first to verify availability, context, supported parameters, pricing, and suff
- Create Response
Create a non-streaming OpenAI Responses API result from plain text or structured multimodal input with a selected ImageRouter text model. Call IMAGEROUTER_LIST_MODELS and IMAGEROUTER_GET_CREDITS first because all text models observed during
- Edit Image
Transform or inpaint one or more existing images supplied as URLs or data URIs. Call IMAGEROUTER_LIST_MODELS and inspect the selected model first to verify image and optional mask input support, accepted parameters, pricing, and funding req
- Generate Image
Generate an image from a text prompt with a selected ImageRouter model. Call IMAGEROUTER_LIST_MODELS and inspect the selected model first because pricing, supported sizes and quality levels, and funding requirements vary. Results are return
- Generate Video
Generate a video from text and optional URL or data-URI image inputs. Call IMAGEROUTER_LIST_MODELS and inspect the selected model first because pricing, accepted modalities, quality levels, sizes, and duration constraints vary. Results are
- Get Credits
Inspect remaining credits, cumulative usage, deposits, and optional per-key usage before making potentially paid generation calls.
- Get Model
Get current capabilities, accepted parameters, and pricing for one exact ImageRouter model ID before generation.
- List Models
Discover current ImageRouter model IDs, modalities, supported parameters, and pricing before generation. Filters may be combined. limit truncates results and cannot be continued; omit it for an exhaustive filtered list.
Setup
Setup guide
- 11. Sign up at ImageRouter and purchase credits, then copy your API key from the dashboard. 2. In Switchy, open your workspace settings and navigate to the Integrations tab. 3. Click 'Add MCP Integration' and select ImageRouter from the catalog. 4. Paste your API key into the authentication field and click 'Connect'. 5. Switchy will validate the key and display your current credit balance. 6. Open any Project and type '@ImageRouter list available image models' to confirm the connection works. 7. The MCP will return a catalog of models with pricing and supported parameters. 8. To generate content, @mention ImageRouter again with a specific request like '@ImageRouter generate an image of a mountain sunset using flux-pro' and the assistant will call the appropriate tool, deducting credits from your balance.
What teammates see: by default, memories from ImageRouter are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Compare Image Styles
@ImageRouter generate 'a cozy coffee shop interior, warm lighting, plants' using flux-schnell, then midjourney-v6, then stable-diffusion-3, and show me all three results side by side
Edit Product Photo
@ImageRouter edit this product photo URL to remove the background clutter using flux-fill-pro, keeping the subject sharp and centered
Generate Explainer Video
@ImageRouter generate a 5-second video from 'a rocket launching at dawn, cinematic angle' using runway-gen3-turbo and confirm the credit cost first
Check Model Pricing
@ImageRouter show me the pricing, supported sizes, and quality levels for midjourney-v6 so I can budget this batch job
Run Advanced Reasoning
@ImageRouter create a chat completion with claude-opus analyzing this financial dataset and explaining three investment scenarios in structured JSON
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@ImageRouter list all available image generation models under $0.05 per image and show their supported resolutions
Found 12 image generation models meeting your criteria. Top options: flux-1-schnell ($0.003/image, supports 256×256 to 1440×1440), stable-diffusion-3-medium ($0.035/image, up to 1024×1024), and playground-v2.5 ($0.008/image, square formats only). Each model has different speed/quality trade-offs — flux-1-schnell generates in ~2 seconds but with less detail than SD3. All models accept standard prompt formatting and negative prompts. This is an illustrative summary; actual model availability and pricing may vary by account tier.
This example demonstrates the MCP's model discovery capability. Before generating any images, you should call List Models to verify current pricing and capabilities — model costs and availability change frequently. The MCP requires API key authentication and sufficient account credits.
@ImageRouter generate a 1024×1024 image of a minimalist workspace setup using flux-1-schnell, then check my remaining credits
Image generated successfully with flux-1-schnell. Cost: $0.003. Image URL: [https://cdn.imagerouter.io/gen/abc123...] (expires in 24 hours). Your account now has 847 credits remaining (approximately $8.47 at current rates). Generation took 2.1 seconds. Note: This URL is temporary — download or reference the image promptly if you need to preserve it. This example illustrates a typical generation workflow; actual output URLs and credit deduction depend on your account state.
This showcases the core generation workflow plus credit tracking. ImageRouter charges per generation, so always check credits first with Get Credits. Generated image URLs expire after 24 hours — the MCP cannot store images permanently. Each model has different cost/speed/quality characteristics.
@ImageRouter take this product photo [URL] and remove the background, then generate three color variations using stable-diffusion-3 inpainting
Background removed from source image. Generated three inpainted variations with color shifts: warm sunset tones ($0.042), cool blue-grey palette ($0.042), and high-contrast black/white ($0.042). Total cost: $0.126. All images maintain the original 1200×800 resolution. Inpainting preserved product edges with 94% accuracy based on automatic mask detection. Output URLs: [cdn.imagerouter.io/edit/xyz789...] (each expires in 24 hours). This illustrates the edit workflow; actual mask quality and color accuracy depend on source image complexity.
This demonstrates the Edit Image tool for transformation tasks. Inpainting and background removal require models that support mask inputs — verify with Get Model first. Each edit operation consumes credits. Complex edits may need manual mask refinement for best results, which this MCP cannot provide.
Use-case deep-dives
When ImageRouter fits a 3-person content team's paid-model budget
A 3-person marketing team ships 20 social posts a week and needs hero images, video clips, and occasional edits. ImageRouter is the right call if you're already paying for multiple AI subscriptions and want one API to route between them based on cost or quality. The Get Credits and List Models tools let you check balances and compare pricing before each generation, which matters when you're burning through 80+ images a month. The trade-off: if your team only uses one model provider (say, just DALL-E or just Midjourney), you don't need the router overhead. But if you're juggling three vendors and want Switchy to pick the cheapest option for each task, ImageRouter consolidates the auth and the decision logic. Buy this if your monthly AI spend is over $200 and you want programmatic cost control.
When image editing beats screenshot annotation for support tickets
A 5-person support team handles 40 tickets a day, many requiring annotated screenshots or redacted customer images. ImageRouter's Edit Image tool (inpainting and transforms) works if your workflow is 'customer sends screenshot, agent marks the problem area, AI generates the fix or redaction'. The API-key auth means each agent can share one Switchy workspace without separate logins. The constraint: this only wins if your support platform doesn't already have built-in annotation. If you're using Zendesk or Intercom with native markup, adding ImageRouter is overkill. But if you're triaging in Slack or Linear and need programmatic edits (blur a credit card, circle a UI bug), the MCP saves you from opening Photoshop 40 times a week. Buy this if your support queue includes visual edits and your current tool requires manual export.
When video generation replaces screen-recording for feature launches
A 4-person product team ships a new feature every two weeks and needs a 15-second demo video for each launch email. ImageRouter's Generate Video tool is the right fit if you're starting from static mockups or wireframes and want AI to animate them, rather than recording your screen. The List Models call shows which video providers (Runway, Pika, etc.) are available and their per-second pricing, so you can budget before generation. The boundary: if your feature is interactive or needs real UI footage, screen recording still wins. But if you're launching an API, a backend change, or a concept that's easier to illustrate than demonstrate, text-to-video cuts your production time from 2 hours to 10 minutes. Buy this if your launch cadence is weekly or faster and your features are hard to screen-capture.
Frequently asked
What does the ImageRouter MCP do in Switchy?
It routes AI generation requests—text, image, video—to multiple paid models through a single API. Your team can call OpenAI-compatible chat completions, generate images from prompts, edit existing images, or create videos without managing separate vendor accounts. Switchy agents check your ImageRouter credits and model availability before each call, so you never hit a surprise quota wall mid-workflow.
Do I need an ImageRouter API key to connect this MCP?
Yes. Switchy authenticates via API key, which you generate in your ImageRouter dashboard. The key must have sufficient credits and permissions to call the models your team wants to use. One key can serve your entire Switchy workspace; you don't need separate keys per user. Paste it into Switchy's MCP settings once and the connection persists across all agents.
Can the ImageRouter MCP generate videos from text prompts?
Yes, using the Generate Video tool. You supply a text prompt and optionally an image URL or data URI; ImageRouter routes the request to a video model like Runway or similar. The MCP checks model capabilities and pricing first, so your agent knows which sizes, quality levels, and durations the selected model supports before spending credits. Output is a video file URL.
How does this compare to calling OpenAI or Midjourney directly?
ImageRouter aggregates dozens of models—text, image, video—behind one API, so you skip juggling multiple vendor keys and billing accounts. The trade-off is an extra routing layer and per-request markup on top of base model costs. Use it when you want model flexibility and unified billing; stick with direct APIs if you only need one or two models and want the lowest possible per-token price.
Who on the team should connect the ImageRouter MCP?
Whoever holds your ImageRouter API key and manages credit top-ups. That person pastes the key into Switchy once; after that, any workspace member can build agents that call ImageRouter models. Monitor the Get Credits tool output regularly—agents will fail silently if your balance hits zero, and ImageRouter doesn't auto-reload unless you configure it in their dashboard.