Imaginode
Chain 74 AI models on one node canvas
About Imaginode
Imaginode is a node canvas for AI creation. Instead of moving files between a dozen separate tools, you place image, video, text and audio models on one board and pipe each output into the next node: a product photo becomes a video, that video gets a voice-over, and the whole chain stays visible and rerunnable. The point is not any single model but the wiring between them, which is the part that usually lives in a folder of intermediate files and a person's memory.
Seventy-four models share one credit balance, which removes the thing that makes multi-model work expensive: a separate subscription per provider, most of them idle most of the month. Named engines include Seedance 2.5, Kling and Veo for video and Seedream 5 Pro for images, alongside ChatGPT, Gemini and Claude for text. Every plan unlocks all of them and only the credit volume changes, so choosing a tier is a question of how much you generate rather than which features you are allowed.
You start a canvas and drop in a node for each step. A style node feeds an image node, which feeds a video node, which feeds a voice-over node. Each node names the model it uses and the credits that call will cost, and the cost is shown before you click rather than discovered afterwards. Run the chain, change one node, run it again: the canvas is a project you can replay rather than a sequence of one-off prompts. Ready-made templates cover product shots, lookbooks, ghost mannequin, comic strips and cinematic ads if you would rather start from a working graph than an empty board.
- •One credit balance across 74 models - Image, video, text and audio models draw on the same balance, so there is no per-provider subscription and no unused seat sitting idle.
- •Cost shown before every generation - Each node displays what the call will spend before you commit to it, which is unusual in credit-based tools and makes an expensive video node obvious in advance.
- •Failed generations are refunded - Credits are returned automatically when a generation fails, so a bad run costs time rather than money.
- •Chained nodes on a replayable canvas - Output pipes into the next node and the whole graph can be rerun after changing one step, with unlimited canvases on every plan.
- •API and hosted MCP server - The same models and the same credits are reachable from Claude, Cursor or your own scripts with an API key, with failures refunded there too.
- •Fourteen language interface - Including French, Spanish, German, Portuguese, Japanese, Korean, Arabic, Hindi, Russian, Turkish and Hebrew.
Suits creators and small studios producing multi-step visual work, where the value is in the chain rather than a single generation: a product shot that becomes a lookbook that becomes an ad. The API and MCP server make it worth a look for developers who want image and video generation inside an agent workflow without integrating four providers separately.
Three things to know before subscribing. There is no free tier, only two free images on signup with no card, so evaluating it properly means paying. Subscription credits reset to the plan allowance at renewal rather than rolling over, though separately purchased top-up credits do not expire and monthly credits are spent first. And paying means accepting immediate access and waiving the fourteen day right of withdrawal, with used credits non-refundable, which is worth reading before a large top-up.
Pricing
$17 - $188/mo
- Starter$17/mo
- Power$55/mo
- Max$188/mo
From the vendor pricing page, 2026-10-08
