Can Claude generate images and videos? Not on its own—here is how to do it with a connector

Claude doesn't generate photos, illustrations or video footage itself, Anthropic's help center says. Add LUVI Creator as a custom connector and Claude can quote and run 200+ image, video, audio and 3D models on your LUVI credits, with the price shown before each run.

Published:

A brass drafting pen draws a glowing teal wireframe chart on black paper; an indigo cable runs from it past an amber bead in a brass dish to a film strip, a landscape photo and a ribbon of sound in warm colour.

Claude does not generate photos, illustrations or video footage on its own. Anthropic's help center says so, and its two built-in visual features produce code, not pixels. What Claude can do is use a connector: add LUVI Creator as a custom connector and Claude picks a model, quotes the price in credits, generates on your LUVI account and shows you the result. That covers 200+ models for video, image, audio and 3D in the LUVI model catalog, including Veo 3.1, Kling 4.0, Seedance 2.5 and Nano Banana. On November 11, 2026, the quote for an 8-second Veo 3.1 Fast clip at 720p was 1,280 credits without sound and 1,536 with it.

Can Claude generate images or video on its own?

No. Anthropic's help center states it directly: "Claude doesn't generate photos or illustrations the way image-generation tools do." What Claude builds itself are diagrams, charts and interactive visuals, made with HTML and SVG.

Video is the same story. Claude Motion makes animations you can export as MP4, but Anthropic's guide says "Claude Motion doesn't use a video generation model, so it doesn't create realistic footage or people." An animated chart is in reach; a filmed-looking shot of a product is not.

Why do you add an image or video generator yourself?

Because Anthropic's own connector directory doesn't take them. The Anthropic Software Directory Policy lists, among the software it does not allow in its directories unless permitted in writing, "Software that uses AI models to generate images, video, or audio content", with an exception for design-focused tools.

Claude still lets you connect one yourself. Anthropic's documentation says you can add any remote MCP server as a custom connector by its URL, on Free, Pro, Max, Ultra, Team and Enterprise plans; the Free plan allows one custom connector. A custom connector is an unverified service, so add only one from a company you trust.

How do I add LUVI to Claude?

You add LUVI Creator as a custom connector with one URL, then sign in with your LUVI account.

  1. In Claude, go to Customize > Connectors.
  2. Select Add, then Add custom connector.
  3. Name it LUVI, enter https://www.luvicreator.com/api/mcp and click Continue.
  4. Keep the authentication options Claude marks Detected, then click Add.
  5. Click Connect. Claude opens a LUVI sign-in page; sign in and approve the connection.

On a Team or Enterprise plan, an Owner adds the connector under Organization settings first, and members then connect with their own account. There is no API key to paste: the approval is tied to your LUVI account. The LUVI help article for Claude has a short video of the steps.

How do I ask Claude to generate?

Name LUVI, the model and what you want, and ask for the quote first.

Using LUVI, quote an 8-second 16:9 clip on Veo 3.1 Fast with sound: a slow push-in on a ceramic mug steaming on a windowsill at dawn, soft window light, the low hum of a kettle and birdsong outside. Generate it only if it costs less than 2,000 credits.

Three things happen before any credit is spent:

  • Claude reads the model's prompting rules. LUVI keeps a prompting guide for each model family. The connector's generate tool stops without charging until the assistant has read the rules for the model it chose, and Claude tells you "LUVI Prompt Engine is active" instead of narrating the steps.
  • The prompt is checked. check_prompt checks the draft against the rules a machine can check, such as syntax, length, and a ratio or resolution written into the text, and returns a corrected version.
  • The price is quoted. estimate_cost returns the credits the run will take, priced the same way the charge is. Settings written in your message, such as "16:9" or "with sound", are moved into the model's parameters.

Then Claude generates, follows the job and shows the result in the chat with the credits it actually cost.

What does it cost, and can I cap it?

Connecting costs nothing; credits are spent only when a generation or a workflow run starts. Prices are in LUVI credits, quoted before the run and reported after it.

  • Before. The quote is free. On November 11, 2026, through the connector: Veo 3.1 Fast, 8 seconds at 720p, 1,280 credits without sound and 1,536 with sound; Kling 4.0, 5 seconds at the default settings, 2,139 credits; Seedance 2.5, 5 seconds at the default settings, 3,049 credits.
  • After. Every finished generation reports the credits it took, and list_transactions returns the same ledger as Account → Billing.
  • Cap. Under Account → Integrations you can set a credit cap for each connection. Once the cap is reached, that connection stops generating.
  • Rules. A connection can also carry rules, such as a credit ceiling per generation or models it may not use.
  • Revoke. Revoking removes the connection at once, and nothing it already made is deleted.

Every model in the catalog runs on every LUVI plan, including the free pay-as-you-go plan, so the model Claude picks is never locked behind a higher tier. The one exception is a small set of permissive models that are opened per account on request.

Where do the results go, and can I use my own files?

Results are saved to your LUVI library, where you can open them in the Workspace, reuse them as references or carry them into a workflow. Claude also shows images and plays audio in the chat and gives you links to the files.

Your own files reach LUVI in two ways. Claude can give you an upload link that opens LUVI's upload window in your browser, or, when Claude can run code, it can upload a file you attached to the chat directly. Images can be up to 50 MB and other files up to 1 GB, as the upload help article explains.

Can Claude chain several models?

Yes. Claude can build a LUVI workflow, a canvas where one model's output becomes the next one's input, for example an image that becomes a video and then gets a voice-over. A wrong connection is refused before anything is saved, and the whole run is estimated before it starts.

Running is your call. Claude can start the run only if you have allowed it on the canvas, with a credit ceiling and an expiry; otherwise it hands you a link and you press Run yourself. How to build a workflow from Claude covers the steps.

Does this work in Claude Code and in ChatGPT?

Yes to both, with a different setup for each.

LUVI is not the only platform with a connector. We compared several of them, from their own documentation, in the MCP connectors compared.

How we checked

  • What Claude does on its own: Anthropic's help center articles on images and on Claude Motion, and its Software Directory Policy, read on November 11, 2026.
  • Setup steps: Anthropic's documentation on adding a connector that isn't in the directory, and LUVI's own help article and connector documentation, read on November 11, 2026.
  • Quotes: four estimate_cost calls through LUVI's connector on November 11, 2026. The Kling 4.0 and Seedance 2.5 numbers match the starting prices on their public model pages.
  • Results: No outputs were generated for this post, so it makes no claim about output quality.
  • Drawbacks: The quotes were made from Claude Code, not from the Claude app. We did not test each Claude plan ourselves; the plan rules are Anthropic's as published on November 11, 2026.

In LUVI

Add the connector once, then work in this order: name the model, ask for the quote, read the number, and let Claude generate. Set a credit cap for the connection before you start if you want a hard limit.

New to LUVI? Create a LUVI account, connect it to Claude, and begin by asking what one clip would cost.

Sources

Questions

Can Claude generate images?

Not photos or illustrations. Anthropic's help center says Claude doesn't generate them the way image-generation tools do; it builds diagrams, charts and interactive visuals in HTML and SVG. With a connector such as LUVI Creator, Claude can run image models on your account.

Can Claude generate videos?

Not footage. Claude Motion builds animations as code and, in Anthropic's words, doesn't use a video generation model. With LUVI Creator connected, Claude can run video models such as Veo 3.1, Kling 4.0 and Seedance 2.5 on your LUVI credits.

Does Claude show the price before it generates through LUVI?

It can. LUVI's estimate_cost tool returns the credits a run will take before anything is spent, and you can set a credit cap for the connection in your LUVI account.

More from Guides