Loading, please wait ...
Connect Vidofy once and your assistant can reach 573 models across 21 modes — describe what you want, see the price first, and get the finished file back in the conversation.
Your connector URL
https://vidofy.ai/mcp-app
New accounts start with 40 free credits — about 40 images to try it with.
Connect from
Claude
ChatGPT
VS Code
Claude Code
Codex
Cursor
Gemini CLI
Hermes
…and any other MCP client — by URL, or by token where the client prefers one.
Three steps in every one of them. Most take the URL and a browser sign-in; a couple take a personal token instead.
Custom connectors are available on every Claude plan, including Free — where you get one.
https://vidofy.ai/mcp-app
You write a sentence. It picks the model, tells you the price, runs it, and hands back the file — in the same conversation.
“Make me a 9:16 clip of a cyclist riding through neon-lit rain, about 5 seconds. Tell me what it costs first.”
list_models
Looks through the image-to-video and text-to-video models for one that does 9:16 at 5 seconds.
get_model
Reads that model’s input contract — which fields it accepts, which are required, what durations are valid.
estimate_cost
Answers the “what does it cost” part BEFORE anything runs. You approve, or ask for a cheaper model.
generate → get_status → get_result
Runs it, follows it while it renders, and returns the finished clip in the conversation.
On Claude and other MCP Apps clients the result arrives as a live card that updates itself while the clip renders — progress, then the finished video, with its cost on it. On a text-only client the same information comes back as text. Nothing is lost; it just looks plainer.
And the result can start the next one — ask for “the same shot but at sunset” and it reuses the file already in your account instead of re-uploading anything.
In plain sentences. You never name a parameter unless you want to.
Point it at an image already in your account and ask for a cleaner background, a different setting, or three variations to choose from.
Hand it a photo and describe the movement. It finds an image-to-video model that accepts your aspect ratio and duration.
Text in, speech out, in the voice and language you name — then use it as the audio for a video in the same conversation.
Ask for the same idea in four aspect ratios, or one scene at three times of day. It runs them and brings them all back.
Any finished result can be the input to the next generation — the file is already in Vidofy, so nothing is downloaded and sent again.
Ask what a plan would cost before running it, and what you have left afterwards. The price comes from the same calculator the studio uses.
Every family below is in the catalogue right now — no separate subscription for each, and no waiting for us to expose a new one. A model added today is reachable from your chat today.
This is the whole surface your assistant is given. Anything not on this list, it cannot do.
Browse what Vidofy can generate and what each model needs, so it picks one that fits the request instead of guessing.
list_modes · list_models · get_model
The exact credit cost before anything runs. Ask “what would this cost?” and you get a number, not a surprise afterwards.
estimate_cost
Video, images, voice and effects — started from the conversation, at the same price as the studio.
generate
Generation takes time. It checks progress and tells you when the file is ready.
get_status
The finished file comes back in the conversation — an image appears as an image, not a link to open.
get_result
Your balance, what expires with your subscription, and what recent generations cost.
get_balance · get_usage
It cannot buy credits, change your subscription, reach your payment method, or see anyone else's media. Purchases happen on this website, by you.
You never copy a credential, never paste one into a config file, and never have one sitting in a chat log. Your client sends you to Vidofy, you approve the connection in your own browser, and the client holds the permission — not you. Revoke it from your account whenever you like and that client disconnects on its next call, while the others keep working.
A connector that surprises you is worse than one you never added. So here is everything that usually comes as a surprise.
The connector is built for conversation. For scripts, CI and terminal agents where you name the model and the flags yourself, the CLI is the right tool.
See the Vidofy CLIEvery finished generation can be the input to another — an image becomes the first frame of a video, a clip gets a voice-over. The file never leaves Vidofy, so nothing is downloaded and uploaded again.
What people ask before they connect it.
No. You paste one URL into your AI client, and the client opens a Vidofy sign-in page in your browser. You approve the connection once and the client is connected from then on. There is no key to copy, store or rotate, and nothing to paste into a config file.
Claude, ChatGPT and VS Code all take the connector URL, and so does any other client that accepts a remote MCP server — Claude Code, Codex and Cursor among them. Claude allows them on every plan, including Free, where you get one connector. In ChatGPT they live under Settings → Connectors, sometimes behind Developer Mode, and on a Business or Enterprise workspace an administrator adds them for everyone. A few clients sign in a different way and cannot use the URL at all — Gemini CLI and Hermes are the two you are most likely to be holding. Those run the npm package on your own machine against a personal token instead, which reaches the same account, the same models and the same balance.
The tools it is given: your credit balance, your recent generations and what they cost, the model catalogue, and the results of generations. It cannot read your email, your password, your payment details or anyone else's media.
It can spend credits you already have, because generating costs credits — and it can tell you the exact price before it runs anything. It cannot buy credits, start or change a subscription, or touch your payment method. Purchases happen on the Vidofy website only.
You can point a generation at media that is already in your Vidofy account, and at a result the assistant just produced, so you can chain one generation into the next. Uploading a file from your computer through the chat is not supported yet — upload it in the studio first.
The connector changes nothing about that: the same rules as the website apply. Subscribers and approved creators get clean files, audio is never watermarked, and the connector tells the assistant which of the two it is looking at rather than guessing.
Yes — one account, one balance. A generation started from your chat costs exactly what the same generation costs in the studio, appears in your studio history, and draws from the same credits.
A year, then your client asks you to approve it again. You can also revoke it at any moment from your account, which disconnects that client immediately without touching the others.
Judge one on three things: how many models it reaches, whether it tells you the price before it spends your money, and whether it needs an API key. Vidofy's connector reaches the whole Vidofy catalogue — Veo, Kling, Sora, Seedance, Nano Banana, Flux, GPT Image and more, across video, image and voice — quotes the exact cost before running anything, and authorises through your browser with no key to manage. One balance covers all of it.
The connector is for a conversation: you ask in words and the assistant chooses the model and the settings. The CLI is for a terminal and for scripts, where you name the model and the flags yourself and the output is predictable enough to put in a pipeline. Same account, same credits, different hands.
One URL, one approval in your browser, no key to manage.
Create a free account
