PixelDojo lets an AI assistant find an effect before paying to run it

“Make this photo into a short clip” is easy to ask and harder to budget. PixelDojo’s October 18 update gives its Sensei assistant, MCP server and API a shared way to find and run named effects. A creator can inspect the available operation and its price before starting a render. October changelog.

Find the operation, then authorize the render

MCP is a connection through which a compatible assistant can call another service’s tools. Here, pixeldojo_effects searches the catalogue and returns each match’s identifier, required inputs and credit price. That lookup consumes no generation credits. pixeldojo_effect executes the selected effect with its photo inputs, returning media or an asynchronous job handoff. Execution uses the listed engine rate and is charged on success. Official tool cards.

A useful proposed request would be: find a product-photo effect, show its required inputs and total quoted credits, and stop before running it. Compare two matches against the actual job. A camera move may suit a product detail; a stylized transformation may change the appearance you need to preserve. Choose one only after seeing the operation and quote. We have not connected or executed this example.

This separates a creative suggestion from a billable action. It also makes an assistant’s vague promise inspectable: a returned job identifier is something to follow to completion, not evidence that a finished clip already exists.

Know which connection and allowance you are using

The hosted MCP route uses browser approval; local-package and REST API routes use an API key. The documentation describes those as different setup paths. Check support in your chosen assistant before authorizing a connection. Connection FAQ.

PixelDojo’s pricing page lists three free video or photo effects per day with a small watermark and no card required. That website offer does not establish unlimited free MCP or API generation. Use the selected tool’s current quote and account allowance, and inspect the output before client delivery. Our AI output evaluation guide can help define that final check.