About Sequelone connection to the whole catalogue

Why it exists

Generative media moved faster than the integrations around it. A team wires an agent to one image model, then a video model arrives that is better at the thing they actually needed, and the work starts again: another account, another billing relationship, another shape of request and another way of waiting for the result.

Sequel is the layer that makes that churn somebody else’s problem. Your agent connects once, describes the job in words, and the catalogue picks the model. When a better one lands, it is already there.

What it is

One MCP server in front of the image, video and audio models worth using, with the plumbing they need around them — a searchable catalogue with schemas, estimates before a run starts, a task queue for the generations that cannot answer inside a request, webhooks for delivery, and a record of who ran what, from which client, on which key.

Everything the agent can do is in the docs, and everything it can run is in the catalogue.

How we work

Small, and close to the thing being built. We would rather ship the capability than the announcement, which is why the catalogue and the pricing page both read from the same registry the server runs on — if it is on the site, it works.

Talk to us

Questions, a model you want added, or something that broke: musthaq@sequel.sh. There is no ticket queue between you and the people building it.