Skip to main content
The same tools run in two places. The cloud means Eversince’s servers, reached through the hosted MCP server, the REST API and the CLI. The desktop means the Eversince app on the user’s Mac or Windows computer, which offers the same three interfaces on that computer. The iPhone app doesn’t run tools. It shows the work and can bring in photos and videos from the phone. Where a call runs depends on where it is sent, not on anything in the call. A call to the hosted interfaces runs on Eversince’s servers. A call through the app runs on the computer whenever the app can do the work there. Every tool has one schema and one set of rules in both places.

In the cloud

Work on Eversince’s servers runs the same from any host, with nothing open on the user’s side. A slow operation returns a job and keeps running after the call ends, whether or not the connection stays open, and get_jobs or a webhook reports the result. Generation and other AI model work is charged in AI credits. Transcribing, reading, describing, rendering and converting are charged in cloud processing minutes, and a response that used minutes includes cloud_processing with what it used and what is left.

On the desktop

The app works on files where they already are. Work that would use cloud processing runs on the computer for free when the app can do it: reading files, transcribing, detecting beats, speakers and loudness, indexing a video, reframing, syncing recordings, making preview frames, and rendering at any length and resolution. Generation and model reads (a prompt on inspect_media, a shot log) run in the cloud wherever the call comes from. Each tool’s description says how it works in the cloud, and on the desktop this rule applies on top of that. A file on disk can be added to the library with import_media in local mode and its absolute path. It takes one call, nothing is uploaded, and the file can be played and edited right away. By default the app also syncs desktop files to the cloud while the owner’s storage has room, so members, phones and cloud agents can use them. The workspace setting sync_files_to_cloud turns that off, and a file then stays on the computer it came from. When a cloud-only read (inspect_media with a prompt or in shot log mode) is called through the app on a video or audio file that is only on that computer, it does nothing and returns what the upload and the work would cost. If the file is only on another computer, the call is refused with that computer’s name. The same call with upload_to_cloud: true uploads the file first, then runs. Any other cloud-only tool refuses and names the computer that holds the file, where the app can sync it to the cloud.

Fields the desktop interface adds

Through the app, these fields appear on the tools that use them. The hosted interfaces don’t list them, since there is no computer there, and a call that sends one is refused with a short explanation.
  • run_on: "cloud" sends the call to Eversince’s servers anyway, for work that has to keep going with the laptop closed.
  • run_on: "local" refuses instead of using cloud processing. The error is needs_cloud, or a refusal saying the computer could not render what was asked, and the work that needed the cloud does not run.
  • upload_to_cloud: true uploads a file that is only on that computer before a cloud-only read runs on it.
  • paths on import_media in local mode: absolute paths on that computer, up to 10 a call.
The desktop app’s own instructions, sent with its tool list, include these rules. Apart from them, an agent connected through the app and one connected to the hosted interfaces see the same tools, descriptions and costs.