Figma Context MCP
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Descripción del creador (original en inglés)
Local video post-production: project inspection, transcription, signed edits and export.
Estos tres nunca se combinan en una sola puntuación: las opiniones, las mediciones y la popularidad son cosas distintas.
Todavía no escribimos un resumen sencillo para este. Abajo están los datos que leímos del registro.
Elige la app de IA que usas y copia la configuración.
¿Qué app usas?
Elige tu app arriba. La recordamos, así que la próxima vez el botón dirá “Instalar en …”.
Descargar el archivo .mcpb del creador
El enlace lleva al archivo del propio creador; no lo alojamos ni lo reempaquetamos. Abre el archivo y Claude Desktop lo instalará.
Ejecuta esto en tu terminal
Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Este se ejecuta en tu computadora, y ChatGPT solo puede conectarse a servidores en internet. Elige otra app.
Tu elección se guarda en una cookie en este dispositivo; no necesitas cuenta.
Esta configuración se genera a partir de la ficha del creador en el registro. Si el README dice otra cosa, sigue el README.
Se ejecuta en tu computadora; la lista de herramientas aparece en tu app de IA después de instalarlo.
Revisamos información que cualquiera puede ver; no hemos leído el código. Las revisiones que no aplican no se cuentan. · Revisado el 7 nov 2026
El nombre en el registro se verificó con la cuenta de GitHub jiyangnan, dueña del repo
Todavía no hemos leído este repo en GitHub
Todavía no hemos leído este repo en GitHub
Este se ejecuta en tu computadora; no hay endpoint en internet
Se ejecuta en tu computadora y usa ajustes locales en su lugar
Se ejecuta en tu computadora; solo podemos leer listas de herramientas de endpoints en internet
Comparado con todos los servidores activos de esta categoría en el registro. Las estrellas solo cuentan cuando quien lo publica es dueño del repo.
Superó 1 de 1 revisiones básicas aplicables. No hemos leído el código. Antes de agregar una clave o un token, revisa qué acceso pide el servidor y usa una clave que tenga solo los permisos que necesitas.
El registro no indica ninguna clave obligatoria, aunque algunos servidores te piden iniciar sesión la primera vez que los usas.
Se ejecuta en tu computadora. Tu app de IA lo inicia cuando hace falta, y los datos no pasan por los servidores del creador (a menos que el propio servidor llame a un servicio en línea).
El repo de GitHub cambió por última vez –, y la última versión (0.2.1) llegó al registro hace 24 días.
Las apps compatibles con MCP, como Claude Desktop, Claude Code, Cursor y VS Code. También necesitas MCPB en tu computadora.
En la misma categoría: Figma Context MCP, Claude Real Video, Figma MCP Server
Servidores de la misma categoría hechos por otros creadores, ordenados por estrellas.
Cómo leer una tarjeta:★ Estrellas de GitHubiTres pilares: reseñas · revisiones · usoiActualizado = fecha en que se publicó la última versión en el registroi
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Let any LLM watch a video locally — and search everything it has ever watched.
The Figma MCP server brings Figma design context directly into your AI workflow.
Copiado tal cual del repo del creador. PopNx MCP no lo ha revisado ni lo respalda. Se quitaron HTML, imágenes y scripts. Los enlaces llevan fuera de este sitio.
AgentMesh-CreatorCut is the public execution product for CreatorCut: an Agent for intelligent post-production of talking-head videos and product screen recordings.
M1 is designed for Chinese, English, and mixed-language recordings. The public client keeps media processing local, renders Director-driven interactive cards across hosts, verifies signed edit plans, previews changes, and applies only declarative reversible operations.
macOS or Ubuntu:
curl -fsSL https://raw.githubusercontent.com/jiyangnan/AgentMesh-CreatorCut/main/scripts/install.sh | bash
Windows PowerShell:
irm https://raw.githubusercontent.com/jiyangnan/AgentMesh-CreatorCut/main/scripts/install.ps1 | iex
The installers set up the pinned Node runtime, Git, FFmpeg, whisper.cpp, multilingual transcription model, secure credential backend, CLI shim, and signed managed-update channel. A successful install immediately starts the guided Agent handoff:
creatorcut onboard
onboard checks the local toolchain and trusted Director configuration, asks for the AgentMesh API key through secure standard input, guides local recording import, starts multilingual transcription, and then returns the exact next action. Agents starting another process must invoke the exact next_process.executable with next_process.argv, next_process.cwd, shell: false, and the allowlisted env_overrides merged into the host environment. Embedded API hosts may pass next_argv directly back to executeCli. OpenClaw must instead pass the exact next_openclaw.exec object to its exec tool. That object always uses the fixed literal command creatorcut __openclaw-bridge; argv is carried only in the tool's structured environment field. For the two non-secret JSON-input steps, follow the returned PTY json-line-v1 contract. API keys never use the bridge and remain a manual, private-terminal input. After that private login, OpenClaw changes only auth login to auth status in the retained non-secret continuation; the returned continuation restores the same project-scoped onboard flow and explicit local dependency overrides. The OpenClaw Skill and CreatorCut CLI must be installed from the same verified release archive. A mismatched older Skill fails closed before project access; until the matching Skill reaches ClawHub, reinstall it from the same verified release archive. next_suggested is display-only and uses placeholders when user input is still required. Run the same command at any time to resume from the first incomplete stage.
When a project was created by v0.2.1, v0.3.0 requires a one-time, explicit metadata adoption before status or mutation commands:
creatorcut project adopt-public \
--project /path/to/project.creatorcut \ --confirm-local
This verifies the existing public history and writes only local authority metadata. It does not invoke the experimental internal-project migration. --confirm-local also confirms that every v0.2.1 CreatorCut process has exited and that no legacy preview, Director request, export, or transcription task is still in progress. Do not open the project concurrently in v0.2.1 after adoption. Completed legacy generated assets, previews, and transcription work remain byte-for-byte in place and outside metadata backup, digest, and rollback deletion. Windows v0.2.1 generated-asset paths are validated without rewriting their stored bytes; moving such a project to a POSIX host before adoption is not a supported path-translation workflow. The project state must be on a local filesystem with same-volume atomic rename and reliable SQLite/OS file locking; FAT/exFAT, network shares, and cloud-synchronized virtual filesystems are not supported project locations in v0.3.
See supported platforms for the exact macOS, Ubuntu, Windows, architecture, and dependency contract.
AgentMesh-CreatorCut M1 Cycle 1–5 is complete and public stable releases are available through the signed managed-install channel. The exact activated version and artifact are recorded by the signed AgentMesh Core policy and release/production-smoke-lock.json. Production cloud admission is open for the Director main feature at the AgentMesh Core authoritative price of 50 credits. This repository contains the frozen public Protocol v1 plus a source-buildable local runtime, Director verification code, cross-platform secure credential adapters, stable JSON CLI, public MCP server, semantic card/text presentation adapters, local media import, Chinese/English/mixed whisper.cpp transcription, local Manifest preview/apply/undo/redo, and resumable export.
v0.3.0 adds digest/generation-bound public mutation recovery and verified v0.2.1 public-project adoption. Internal-project migration and rollback remain disabled until their native packaging, installation, platform, and recovery gates are complete.
Batch 3 now includes the independent creatorcut-operations/1.0 contract, immutable *_ref resolution, a real Server-generated signed remove_range fixture, the complete advertised-operation test matrix, Codex MCP App code, the public OpenClaw Skill, generic-text fixture submission, and named session/Generation/preview/export recovery tests.
Cycle 3 is closed for the M1 support matrix: the real Codex native presentation and submit gate, the real OpenClaw Skill run, generic-text fallback, and the four-point recovery matrix passed with one fixed fixture and AnswerSet digest. Claude Code is deferred to a post-M1 compatibility iteration and is not an M1 supported or release-tested host. Its source preview remains under skills/claude-code but is excluded from support claims.
Cycle 4 added the managed installer and signed release chain. Cycle 5 completed the real GitHub RC/stable tag and canonical archive, staging signed Manifest, original-color three-material dogfood, production dark deployment, approved 50-credit paid canary/rollback, persistent Core stable ReleaseManifest and a formal clean public-channel reinstall with real Server/Core preflight.
The public repository is the release carrier, while availability, admission, and the authoritative 50-credit Director price remain controlled by AgentMesh Core.
AgentMesh-CreatorCut (public, local execution)
-> creatorcut-server (private Director strategy) -> agentmesh-core (identity, entitlements, price and credits)
CreatorCut is the internal Studio/dogfood source. creatorcut-server owns card orchestration, semantic editing policy, signed plan generation, and billing recovery. agentmesh-core remains the shared control plane and price source. This product does not depend on AgentMesh360-Client.
See System boundary for the data and trust model.
Requirements:
pnpm install
pnpm verify pnpm protocol:digest pnpm pack:protocol
pnpm protocol:digest prints the stable digest that private consumers must pin. pnpm pack:protocol builds the independently consumable protocol tarball used by private services during this pre-release phase.
After a source build, the development CLI can be invoked with Node 24:
node apps/cli/dist/src/main.js doctor
The local-only media path is available for development verification:
node apps/cli/dist/src/main.js media import \
--project /path/to/demo.creatorcut \ --source /path/to/recording.mov
node apps/cli/dist/src/main.js transcribe start \ --project /path/to/demo.creatorcut \ --model /path/to/ggml-model.bin \ --language mixed
node apps/cli/dist/src/main.js export start \ --project /path/to/demo.creatorcut \ --output /path/to/demo.mp4
zh, en, auto, and mixed transcription modes are supported. The mixed mode runs local auto, Chinese, and English candidates and selects the best bilingual result. Local editing keeps original audio unless a signed Manifest explicitly requests a supported change. Applying a signed Manifest requires the exact confirmation token returned by edit preview; exporting never overwrites an existing file unless --confirm-overwrite is present.
Remote Director commands intentionally require an AgentMesh API key stored by creatorcut auth login in macOS Keychain, Windows DPAPI, or Linux Secret Service; a trusted recovery-root-signed Director keyset; an explicit endpoint, and the pinned protocol bundle digest. The official managed installer provisions the pinned public production trust bundle. A source build does not ship credentials, silently select staging, or bypass that trust configuration.
For the standalone local MCP bundle, see MCP distribution.
See Agent onboarding, privacy contract, and managed release policy. Exact Cycle 4 scope and verification are recorded in Cycle 4 closeout evidence.
This product is part of the AgentMesh360 product matrix — one account powers multiple vertical AI agents.
| Product | Repository | Site |
|---|---|---|
| AgentMesh-JobAgent — AI job-search agent for Chinese recruiting platforms | jiyangnan/AgentMesh-JobAgent | jobagent.agentmesh360.com |
| AgentMesh-CreatorCut — AI post-production for talking-head and product recordings | jiyangnan/AgentMesh-CreatorCut | creatorcut.agentmesh360.com |
| AgentMesh-Lecturecast — agentic course-video production | jiyangnan/AgentMesh-Lecturecast | lecturecast.agentmesh360.com |
| AgentMesh-Runtime — local AI agent memory and recovery | jiyangnan/AgentMesh-Runtime | runtime.agentmesh360.com |
| AgentMesh-Deploy-Agent — open-source deployment operations agent | jiyangnan/AgentMesh-Deploy-Agent | deploy.agentmesh360.com |
| AgentMesh-OfficialRecruitment — official career-site application workspace | jiyangnan/AgentMesh-OfficialRecruitmentAgent | recruit.agentmesh360.com |
Apache-2.0. See LICENSE.