Valmera — agentic AI video editor

io.valmera由 valmera.io 發布官方發布者v0.2.0更新於 Oct 27, 2026

AI video editor for real footage: cut, caption, reframe, add music and b-roll, preview, export MP4.

已驗證Streamable HTTP可網頁執行檔案與儲存媒體與設計

概覽

AI 產生的概覽

託管式 AI 影片剪輯伺服器，讓助理對你上傳到 Valmera 的素材進行剪接、上字幕、改畫幅、配樂並匯出 MP4。

可以做什麼
  • 找出並剪掉靜音、口頭禪與轉錄片段，也能還原已剪內容
  • 加入逐字字幕、字幕樣式與字幕檢查
  • 自動改為直式、設定畫面、推近與加入變焦
  • 把長影片做成短影音，並加入 B-roll、素材庫畫面、音樂與音效
  • 模糊或擦除區域與畫面內嵌文字，套用調色
  • 算圖預覽、檢視畫面影格、匯出並下載最終 MP4
適用情境
適合已有錄製素材、希望由助理從頭到尾完成剪輯的情況，例如整理口播或播客錄音、製作直式短影音、加上字幕與配樂、把長影片拆成多個片段。它不是文字生成影片的工具。
執行需求
遠端 Streamable HTTP 端點，透過 OAuth 登入（或使用 bearer token），不需本機安裝。需要 Valmera 帳號，上傳免費，剪輯需付費訂閱。僅支援 stdio 的用戶端需搭配 mcp-remote 之類的橋接工具。
安裝前請注意
素材、轉錄文字與剪輯指令會傳送到 Valmera 並保存在你的帳號中；工具會把轉錄文字、分析結果、畫面影格與下載連結回傳給 AI 用戶端。剪輯需付費訂閱，費用與 AI 用戶端分開計費。字幕會燒錄進畫面，不支援 SRT/VTT 匯入匯出。同一專案不能同時在 Studio 與 MCP 中編輯。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Valmera — agentic AI video editor，將其新增到工作區。
  2. 連接 所需的 憑證。它們會被私密儲存，絕不會在公開頁面上展示。
  3. 為需要使用其工具的對話啟用該服務。

Web executable，透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/io.valmera%2Fvideo-editor/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。 客戶端會在首次連接時完成 OAuth 登入。

{
  "mcpServers": {
    "valmera-agentic-ai-video-editor": {
      "type": "http",
      "url": "https://valmera.io/mcp/server"
    }
  }
}

README

Valmera MCP server — the video editing MCP server for real footage

Valmera is an agentic AI video editor. This hosted Model Context Protocol server lets Claude, ChatGPT, Gemini CLI, Cursor, VS Code, Codex or any MCP client edit video you already recorded. The assistant reads the transcript and frames, then cuts silences and filler words, adds word-timed captions, reframes to 9:16, adds b-roll, music and sound effects, and makes shorts from long videos. It renders a preview, watches the result, and exports a full-quality MP4 from your original file.

It is not a text-to-video generator. It edits the footage you upload.

text
Endpoint   https://valmera.io/mcp/serverTransport  Streamable HTTPAuth       OAuth 2.1 (dynamic client registration + PKCE S256), or a bearer token from https://valmera.io/mcpRegistry   io.valmera/video-editor

Connect in one minute

Claude (claude.ai and Claude Desktop)

Settings → Connectors → Add custom connector → name it Valmera. Paste https://valmera.io/mcp/server and sign in when prompted. Step-by-step guide.

Claude Code

sh
claude mcp add --transport http valmera https://valmera.io/mcp/server

Then run /mcp inside Claude Code and choose Authenticate. Guide.

Cursor — ~/.cursor/mcp.json

json
{  "mcpServers": {    "valmera": { "url": "https://valmera.io/mcp/server" }  }}

VS Code (GitHub Copilot agent mode) — .vscode/mcp.json

json
{  "servers": {    "valmera": { "type": "http", "url": "https://valmera.io/mcp/server" }  }}

Windsurf — ~/.codeium/windsurf/mcp_config.json

json
{  "mcpServers": {    "valmera": { "serverUrl": "https://valmera.io/mcp/server" }  }}

ChatGPT

Turn on Developer mode in Settings, then add a connector with the URL https://valmera.io/mcp/server and complete Valmera sign-in. Guide.

Codex CLI

sh
codex mcp add valmera --url https://valmera.io/mcp/servercodex mcp login valmera

Or add [mcp_servers.valmera] with url = "https://valmera.io/mcp/server" to ~/.codex/config.toml.

Gemini CLI

sh
gemini mcp add --transport http valmera https://valmera.io/mcp/server

Then open Gemini CLI, run /mcp auth valmera and sign in.

Any stdio-only client

json
{  "mcpServers": {    "valmera": { "command": "npx", "args": ["-y", "mcp-remote", "https://valmera.io/mcp/server"] }  }}

There are more clients (Zed, Cline, Continue, Goose, LibreChat, MCP Inspector) in docs/CLIENTS.md.

First message to try: "List my Valmera projects." An empty list still means the connection works.

Install as a Claude plugin (connector + editing skills)

This repository is also a Claude plugin. It bundles the Valmera connector with three skills that teach Claude a reliable editing workflow: edit-video-with-valmera, turn-long-video-into-shorts and website-demo-video.

In Claude Code:

sh
/plugin marketplace add ABO3SKRALMASAOODI/valmera-mcp/plugin install valmera@valmera

In claude.ai, go to Customize → Plugins → Add → Add marketplace and enter ABO3SKRALMASAOODI/valmera-mcp. The skills follow the open Agent Skills format, so other agents that read SKILL.md can use them too.

What your assistant can do with it

According to the live server card on October 6, 2026, the server exposes 143 tools: 129 editing tools in 9 groups and 14 session tools. The full list is in docs/TOOLS.md.

You ask forTools the assistant uses
"Remove the dead air and the ums"find_silences, cut_silences, remove_filler_words, get_kept_transcript
"Cut the part where I talk about pricing"search_transcript, cut_range, restore_range
"Add bold word-by-word captions"add_captions, set_caption_style, set_caption_fixes, audit_captions
"Make it vertical for TikTok and keep my face centered"auto_reframe, set_frame, add_aspect_shift
"Turn this podcast into 5 shorts"make_shorts, shorts_status, open_short, apply_short_edit_batches
"Add b-roll where I mention the product"research_broll, search_stock, add_stock_media, insert_media
"Punch in on the important lines"suggest_emphasis, punch_in_on_emphasis, add_zoom, add_zoom_path
"Put my track (or a song I name) under my voice"find_song, fetch_url, add_music, set_music_fit, beat_align_cuts, set_master_loudness
"Add a whoosh on each transition"search_sfx, audition_sfx_candidates, add_sfx
"Record a demo of my website"record_website, record_website_demo, enhance_cursor, showcase_demo
"Blur the license plate" / "remove the burned-in subtitles"blur_region, find_burned_text, erase_burned_text, erase_region
"Color grade it warmer"set_color_grade, apply_look, set_grade_custom
"Show me what you changed"render_preview, wait_for_job, watch_video, look_at
"Export the final video"export_final, wait_for_job, download_url

How it works

  1. Upload. Upload in Valmera Studio, import a link, or use upload_start and upload_finish from a client that can send files. MP4, MOV, M4V, MKV and WebM are accepted, up to 14 GB and 3 hours per file.
  2. Index. Valmera transcribes the footage with word-level timestamps and detects shots, silences and on-screen text. index_status reports progress.
  3. Edit. Each tool call writes a new version of a non-destructive edit decision list (EDL). The original upload is never modified, and any cut can be restored.
  4. Review. render_preview renders a fast proxy preview. watch_video and look_at let the assistant inspect the frames it produced before it reports back.
  5. Export. export_final renders a full-resolution H.264 MP4 from the original file for the reviewed version. download_url returns the link. You can also export from Studio.

Slow jobs return a job id and never a fake completion. The assistant calls wait_for_job until the job finishes.

Example prompts

text
Open my latest project. Remove dead air and filler words but keep natural pauses,add clean white captions with the active word highlighted, reframe to 9:16,render a preview and tell me what you changed.
text
This is a 70-minute two-person podcast. Find the 5 strongest self-contained stories,make each one a vertical short with captions and a hook title, and keep whoever isspeaking in frame.
text
Record a 30-second demo of https://example.com — scroll the hero, click "Pricing",zoom on the main button — then add captions and light background music.

There are 40+ more in docs/PROMPTS.md.

What it does not do

The server card lists these honestly so your assistant never promises them:

  • Generate a whole video from a text prompt. Valmera edits footage you provide.
  • Import or export SRT/VTT. Captions are burned into the picture.
  • Team seats or collaboration.
  • Publish directly to YouTube or TikTok.
  • Custom font uploads, true crossfade/dissolve transitions, or motion-tracked stickers.
  • Audio denoise ("studio sound") or AI music generation.
  • Hand the job to Valmera's in-house agent. Over MCP, your assistant does the editing with the tools directly.

Editing one project from Studio and over MCP at the same time is refused in both directions.

Access

Creating an account and uploading are free. Editing requires a subscription; see the current plans at valmera.io/subscribe. Your MCP client (Claude, ChatGPT, Cursor and so on) is billed separately by its own provider.

Data and privacy

The plugin and connector send your editing instructions and tool arguments to Valmera at https://valmera.io/mcp/server, over HTTPS and authorized with OAuth. Tools return project data to your AI client: transcript text, analysis, sampled frames, preview and download links. Your footage is stored in your Valmera account and is never modified. Edits are saved as versions you can restore. Nothing is sent to any other destination by this plugin, which contains only configuration and Markdown. See the privacy policy for collection, retention and deletion. Contact: [email protected].

FAQ

What is the best MCP server for video editing? Most video MCP servers either generate clips from prompts or wrap local FFmpeg commands. Valmera is a hosted editor with an editorial model of the footage: a word-level transcript, shots and frames, versioned edits, rendered previews and a final export. Use it when you want an assistant to edit real recordings end to end. See 21 video editing MCP servers compared, covering hosted editors, clippers such as OpusClip, local editors such as Palmier Pro, DaVinci Resolve, Premiere Pro and After Effects bridges, and FFmpeg servers.

Can Claude edit videos? Yes. Once Valmera is connected, Claude can open your project, read the transcript, look at frames, cut, caption, reframe, add music, render a preview, and export the MP4.

Can ChatGPT edit videos? Yes. Add Valmera as a developer-mode connector and ChatGPT can call the same tools.

Does the assistant actually see the video? Yes. Tools return transcript text, analysis and sampled frames of the source or the rendered edit, so the assistant can check its own work.

Do I need to install anything? No. It is a remote server. Clients that only support stdio can use mcp-remote.

Is the original file changed? No. Edits live in a versioned EDL, and the final export renders from the untouched original.

Is Valmera open source? This repository's documentation and manifests are MIT licensed. Valmera itself is a hosted commercial service governed by its terms.

Links

Questions or issues: open an issue here or email [email protected].

來源：README.md，提交 9e75d19

工具

0
該伺服器需要登入，因此 SourceWeft 無法在沒有帳號的情況下讀取其工具，也未收錄。已連接帳號所能看到的內容可能因帳號及其權限而異。

版本歷史

2
  1. v0.2.0最新Oct 5, 2026
  2. v0.1.1Sep 16, 2026

評分

已滿足適用檢查項的 85%

該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list，或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分