Clipwright
# Clipwright clients
Open-source clients for [Clipwright](https://clipwright.io/?ref=github): an API that turns a
script into a short video of an actor speaking it, or a script or a short brief
into a faceless video without an on-camera presenter: narration over an opening
animated clip and image scenes, ending with the narration (select 30–90 seconds;
it runs 25 seconds or more, usually at most a quarter past the selection and 90; a longer narration runs to its end). Your agent quotes the price for free,
starts the render, and gets a video URL back. Clipwright returns the
file; it does not publish anywhere.
| Package | What it is |
|---|---|
| [`@clipwright/mcp-server`](https://www.npmjs.com/package/@clipwright/mcp-server) | MCP server over stdio for Claude Code, Claude Desktop, Cursor, Cline and any other MCP host |
| [`@clipwright/cli`](https://www.npmjs.com/package/@clipwright/cli) | command-line client (`clipwright`, `clipwright-cli`) |
| [`@clipwright/sdk`](https://www.npmjs.com/package/@clipwright/sdk) | TypeScript client |
| [`@clipwright/core`](https://www.npmjs.com/package/@clipwright/core) | shared request and response schemas (zod) |
## Claude Code plugin
```bash
claude plugin marketplace add seocombat/clipwright-mcp
claude plugin install clipwright@clipwright
```
The plugin connects the hosted server at `https://mcp.clipwright.io/mcp` and
adds a skill that tells the agent to quote before it spends credits. Run `/mcp`
in a session, choose `clipwright` and sign in with your Clipwright account: no
key goes into a config file. The same repository is a Cursor plugin
(`.cursor-plugin/plugin.json`, `mcp.json`).
## n8n
n8n's built-in MCP Client node reaches the hosted server with a key; no community
node is needed. [`n8n/clipwright-ugc.json`](n8n/clipwright-ugc.json) is a workflow
to import: it quotes a script, stops if the quote is above `max_credits`, starts
the render and polls `get_run` every 15 seconds until the run ends. The whole
workflow ran on n8n Cloud on 2026-10-20 and returned the MP4; n8n's OAuth
credential has not been tried with Clipwright yet. Setup:
[`packages/mcp-server/README.md`](packages/mcp-server/README.md#n8n).
## MCP server in one line
```bash
claude mcp add clipwright -e CLIPWRIGHT_API_KEY=cw_... -- npx -y -p @clipwright/mcp-server clipwright-mcp
```
Put your own key from [app.clipwright.io/api-keys](https://app.clipwright.io/api-keys)
in place of `cw_...`. It needs Node.js 20 or newer. Other hosts, the tool list and
the environment variables: [`packages/mcp-server/README.md`](packages/mcp-server/README.md).
Quotes, voices, actors and the balance are free. Renders spend credits: an
account opened by signing up starts with 420 trial credits, enough for one
90-second faceless video. Buy more at
[clipwright.io/pricing](https://clipwright.io/pricing?ref=github).
## Build from source
```bash
pnpm install
pnpm -r build
```
This repository mirrors the client packages of the Clipwright monorepo; each
release is synced here from there. Documentation: [clipwright.io/docs](https://clipwright.io/docs?ref=github).
Privacy: [clipwright.io/privacy](https://clipwright.io/privacy).
MIT License — see [LICENSE](LICENSE).
TDQS
Scored across 57 tools
The dominant pattern — a free quote_X twin for every paid action (quote_ugc/make_ugc, quote_image/generate_image, quote_revision/revise_run) — is stated explicitly in each description, so price-vs-execute confusion is largely prevented, and resources are well separated (runs, assets, uploads, plans, actors). A few genuine boundary cases remain: upload_image vs create_upload (local vs remote server), create_asset's overlapping upload sources, and get_run vs get_run_shots vs get_reference_analysis.
Almost everything is snake_case verb_noun (get_/list_/create_/delete_/quote_/analyze_/approve_), which is highly predictable. Minor deviations exist: the noun-only 'brand_kit', the domain verbs 'make_ugc'/'make_faceless' versus 'generate_image'/'generate_video', and the revise_run/quote_revision mismatch.
57 tools is far past the point where an agent can hold the surface in mind, and much of the volume is structural duplication from the quote/execute/attempt triplet per operation. The domain is genuinely broad (generation, plans, revisions, candidates, branding, actors, trends), but several families (quote+create+revise+candidate for shots) could be consolidated.
Coverage is unusually thorough: quotes, generation, assets/uploads, planning and approval, runs, revisions, candidates, actors, voices, moves, brand extraction, account/credits, plus describe_tool and feedback escape hatches. Gaps are modest — no run listing or cancellation, no deletion of saved workflow plans/revisions, and discovery of past analyses/creators relies on passing the original link.