Skip to main content
D

connectors (deepseek-harness)

dotv-ai/connectors/bundles/deepseek-harness

Drive the DOTV infinite canvas from DeepSeek Harness: the agent lists your canvases and their nodes, adds text nodes, and runs image/video/audio generations — over DOTV's public API, with no runtime dependencies and no build step. It reuses the `dotv` CLI login on the machine, or takes an API key in the plugin settings. Generations consume credits on your account, exactly as they do in the app.

Install

dsh plugin --profile web add github:dotv-ai/connectors

This plugin lives in the bundles/deepseek-harness subdirectory of the repository.

README

DOTV Connectors

DOTV is an AI video creation canvas: one infinite board where a story becomes a script, a storyboard, images, video, voice and a finished cut.

DOTV Connectors let your coding agent work on that canvas. Ask it to turn a script into a storyboard, add a node, run a generation, or just tell you what is already on the board — the work happens on your canvas, and you watch it happen while it runs.

What the agent can do

You sayIt does
"What's on my canvas?"Lists the nodes and what each one holds
"Turn this script into a storyboard"Adds script and shot nodes, wires them into a flow
"Add an image node and generate it"Adds the node, submits the generation
"Add this image to node 12"Puts an existing asset onto the canvas
"Is the video ready?"Waits for the running generations and reports their state

It works with the canvas you already have: text, group, image, video, audio, script, composite and director nodes, connected by edges into a flow.

Install

Codex

codex plugin marketplace add dotv-ai/connectors --ref main
codex plugin add dotv-canvas@connectors

Claude Code

/plugin marketplace add dotv-ai/connectors
/plugin install dotv-canvas@connectors

Any MCP client

claude mcp add dotv -- npx -y @dotv/cli mcp
codex mcp add dotv -- npx -y @dotv/cli mcp

DeepSeek Harness

dsh plugin --profile demo add https://github.com/dotv-ai/connectors/releases/latest/download/dotv-dsh-canvas.tgz

The npm route works too, and needs no build approval either:

dsh plugin --profile demo add @dotv/dsh-canvas

Releases are cut by CI — see CONTRIBUTING.md for how.

What you need

  • A DOTV account — https://dotv.ai.
  • If the DOTV desktop app is open on this machine, the agent's commands run inside it and you watch them happen. Otherwise they run against the cloud and show up the next time you open the canvas.

Running a generation — an image, a video clip, an audio take — consumes credits on your account, exactly as it does in the app.

License

MIT covers the code in this repository (LICENSE); the DOTV product itself is closed source and is not covered by it. Trademarks, credits and the boundary between the two are spelled out in NOTICE.md.

For how this repository is put together and maintained, see CONTRIBUTING.md.

Related plugins