Skip to main content
Glama
Sign In

550W AI去字幕去水印

by cn.550wai
OverviewDiscussionsAdmin

Server Details

图片去文字去水印、本地视频全屏去硬字幕及画面水印，支持抖音、快手、哔哩哔哩、微博等分享链接去平台水印。OAuth 授权；本地文件需宿主上传支持，否则到 qzm.550wai.cn 上传。

Ownership verified
Status
Healthy
Uptime
91.3% over the last 24 hours
OAuth
Works in Glama
Last Tested
Transport
Streamable HTTP · MCP 2025-11-25
URL
ChatGPT
–
Claude
–

TDQS

B3.2/5.0

Scored across 17 tools

Disambiguation2/5

There are parallel, overlapping surfaces: list_image_watermark_tasks vs list_subtitle_tasks vs list_media_tasks (which already covers 'image or video tasks'), and get_image_watermark_task vs get_subtitle_task vs get_media_operation. Two submit tools (submit_subtitle_task, submit_media_task) and two upload tools (prepare_media_upload, prepare_verified_video_upload) target nearly the same intent — removing subtitles/watermarks from video — with the distinction buried in prose rather than made obvious.

Naming Consistency4/5

All names are consistent snake_case verb_noun (get_/list_/delete_/submit_/prepare_/quote_/estimate_/query_), so the pattern is predictable. Minor deviations like the singular noun 'query_credits' and the mixed domain vocabulary (subtitle vs media vs image_watermark) slightly reduce clarity but not pattern consistency.

Tool Count3/5

17 tools is on the heavy side for what is essentially two/three synchronous media-processing pipelines. Each tool has a plausible role in the quote/upload/submit lifecycle, but the duplication of list/get/submit/upload across 'subtitle', 'image', and 'media' domains inflates the surface more than the domain warrants.

Completeness3/5

The subtitle pipeline is well-covered end-to-end (estimate → submit → get → list → delete) and credits/operation receipts exist. However, the generic media pipeline has no delete/cancel counterpart for submit_media_task, and the image path relies on upload auto-submitting a paid task with no explicit create/cancel, leaving some dead ends for recovery and housekeeping.

Available Tools

17 tools
delete_image_watermark_taskB
DestructiveIdempotent
Inspect

Delete your image task using the existing account policy. Deletion is not cancellation or refund.

Inputs: taskId (required).

delete_subtitle_taskB
Destructive
Inspect

删除本人极速去字幕任务；删除不会退还已扣积分

Inputs: taskId (required).

estimate_subtitle_costA
Read-onlyIdempotent
Inspect

提交去字幕任务前，根据视频宽、高和时长估算积分并查询当前余额；实际提交时重新校验。

Inputs: width (required), height (required), duration (required).

get_image_watermark_taskB
Read-onlyIdempotent
Inspect

查询本人图片去水印任务状态、结果和失败原因

Inputs: taskId (required).

get_media_capabilitiesA
Read-onlyIdempotent
Inspect

Discover regional business capabilities, action scopes and input limits; this does not grant access.

Inputs: no named parameters.

get_media_operationB
Read-onlyIdempotent
Inspect

Read an existing HTTP media operation receipt without resubmitting. A missing receipt is not proof of no charge. Namespaces preserve their original keys; old operations lacking region evidence remain unknown.

Inputs: namespace (required), operationId (required).

get_subtitle_taskA
Read-onlyIdempotent
Inspect

查询本人极速去字幕任务的状态、结果和退款状态

Inputs: taskId (required).

list_image_watermark_tasksC
Read-onlyIdempotent
Inspect

List your account image tasks. Page numbering starts at 1.

Inputs: page, size, order, status.

list_media_tasksB
Read-onlyIdempotent
Inspect

List your image or video tasks using the shared account filters. Page numbering starts at 1.

Inputs: kind (required), page, size, order, +1 more.

list_subtitle_tasksB
Read-onlyIdempotent
Inspect

分页查询本人极速去字幕任务

Inputs: page, size.

prepare_media_uploadAInspect

为本地视频或图片申请限时上传凭证。将文件以 multipart 发送到 uploadUrl，凭证放在 X-550W-Upload-Ticket 请求头，无需获取或发送 OAuth Token。图片上传会直接提交付费图片擦除任务，返回 taskId 后用 get_image_watermark_task 查询终态，无需单独创建工具；视频上传仅准备素材，随后调用 submit_subtitle_task。须先获得用户的上传及付费处理同意。图片还必须附带 8–64 字符 operationId，重试沿用相同 ID 和文件；失败、删除、过期不重新创建任务。

Inputs: fileSize (required), mediaType (required).

prepare_verified_video_uploadAInspect

Prepare a generic video upload ticket for verified quotes (up to 1800 seconds, 1080P, 1 GiB; mp4/mov/m4v). Upload only prepares material; then use quote_media_task and submit_media_task. No VOD credentials are exposed.

Inputs: fileSize (required).

query_creditsB
Read-onlyIdempotent
Inspect

查询账户当前积分余额

Inputs: no named parameters.

quote_media_taskB
Read-only
Inspect

Verify a quote for a server-uploaded video mediaId. No charge; not a balance reservation.

Inputs: mediaId (required), area, mode, fileName, +1 more.

remove_video_watermarkA
Idempotent
Inspect

解析从抖音、快手、哔哩哔哩、微博等 App 或网站复制的视频分享链接，返回 data.video 无平台水印直链。本工具不擦除视频画面中的硬字幕或水印；处理本地视频请上传后调用 submit_subtitle_task。客户端下载失败或超时时，把该直链交给用户在浏览器下载。解析成功扣 1 积分；须提供 operationId。

Inputs: videoUrl (required), operationId (required).

submit_media_taskA
Idempotent
Inspect

Confirm a verified quote with a stable clientRequestId. Obtain user approval first; query oauth_v2 operation recovery after an uncertain response.

Inputs: quoteId (required), clientRequestId (required).

submit_subtitle_taskA
Idempotent
Inspect

提交视频去除硬字幕或画面水印。默认整帧擦除；仅在用户明确提供区域时填写全部 x1、y1、x2、y2。按时长和分辨率计费；超时后用原幂等键及原参数重试。

Inputs: width (required), height (required), duration (required), idempotencyKey (required), +10 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 9 tool updates
    • Addeddelete_image_watermark_task
    • Addedget_media_capabilities
    • Addedget_media_operation
    • Addedlist_image_watermark_tasks
    • Addedlist_media_tasks
    • Addedprepare_verified_video_upload
    • Addedquote_media_task
    • Addedsubmit_media_task
    • Changedsubmit_subtitle_task1 field changed
      • addedInput schema / properties / mode
        Added value: +{
+  "enum": [
+    "normal",
+    "protect"
+  ],
+  "type": "string"
+}
  2. 9 tool updates
    • First observeddelete_subtitle_task
    • First observedestimate_subtitle_cost
    • First observedget_image_watermark_task
    • First observedget_subtitle_task
    • First observedlist_subtitle_tasks
    • First observedprepare_media_upload
    • First observedquery_credits
    • First observedremove_video_watermark
    • First observedsubmit_subtitle_task

Related MCP Connectors

Related MCP Servers

  • Short Video MCP Server

    yangbuyiya
    A
    license
    A
    quality
    D
    maintenance
    A FastMCP-based service that extracts no-watermark links from 20+ video platforms (including TikTok, Kuaishou, etc.) and can convert video speech to text.
    3
    65
    MIT

  • Vision OCR MCP

    wbwbwb123
    F
    license
    Not graded
    quality
    D
    maintenance
    Local OCR (RapidOCR) plus optional Qwen-VL for image summarization, supporting both file paths and clipboard images.
    -
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources