Replicate ExtendedをChatGPTとClaudeに接続

• 01 / 03モデルのファインチューニング方法
• 02 / 03デプロイメントの管理方法
• 03 / 03予測を実行する

最速日

789ms

最遅日

1347ms

14日間の傾向

遅延傾向+16%

Replicate Connector

準備完了です。MCPクライアントを選択し、セットアップ手順に従ってください。

コネクタリンクhttps://edge.vinkius.com/vk_preview_ULeVbcRjoLCkIr9x93q2HML8VJd4AXmk4CWHHPso/mcpコピー

Claude Desktop

以下の手順で数秒で接続できます。

1. 1In Claude Desktop, open Settings → Connectors.
2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
3. 3Click Add and start a new chat — Replicate capabilities are ready to use.

設定 · claude_desktop_config.jsonコピー

{
  "mcpServers": {
    "replicate-extended-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_ULeVbcRjoLCkIr9x93q2HML8VJd4AXmk4CWHHPso/mcp"
    }
  }
}

• Claude
• ChatGPT
• Cursor
• VS Code
• Windsurf
• Claude Code
• JetBrains
• Cline

各クライアントの手順ガイドはこちらです。 接続方法

Claude向けのガイド付きセットアップ？ ReplicateをClaudeや他のAIに設定する

このConnectorが対応するすべてのAIクライアントを見る ↑

• 01

  AIエンジニア

  ボイラープレートコードを書くことなく、モデルのバージョンや入力を迅速にテストできます。

• 02

  クリエイティブ・テクノロジスト

  制作ワークフローの中で、直接メディア資産を生成できます。

• 03

  データサイエンティスト

  統合されたインターフェースから、トレーニングジョブや予測出力を一括で監視できます。

• 01

  How do I connect my Replicate account?

  After subscribing to the MCP on Vinkius, you just need to provide your Replicate API Token to establish the connection.

• 02

  Which AI clients can I use with this MCP?

  You can use this MCP with any compatible client, including Claude, Cursor, and Windsurf.

• 03

  Can I run my own fine-tuned models?

  Yes, you can use the training capabilities to create fine-tuned models and then run predictions against them.

• 04

  Do I need to write code to run a prediction?

  No, you can trigger predictions and manage models using natural language commands through your AI agent.

• 05

  Can I manage my deployments through this MCP?

  Yes, you can create new private deployments and manage hardware configurations directly.

• 06

  How can I check if my prediction has finished and see the output?

  Use the get_prediction capability with your Prediction ID. It will return the current status (starting, processing, succeeded, or failed) along with the output URLs or data once completed.

• 07

  Can I search for specific types of models like 'image-to-text'?

  Yes! Use the search_models capability with your query. It will return a list of public models matching your terms, including their owners and descriptions.

• 08

  Is it possible to stop a model that is taking too long to run?

  Absolutely. Use the cancel_prediction capability with the target Prediction ID to immediately stop the execution and prevent further usage costs.

• 

  Modelbit (ML Model Deployments) AIコネクタ

  Deploy and call machine learning models directly from your AI agent using Modelbit's inference endpoints.

  表示
• 

  Replicate AIコネクタ

  Equip your AI to dynamically search, run, and monitor thousands of open-source machine learning models hosted

  表示
• 

  Replicate AIコネクタ

  Run ML models via Replicate — generate images, text, audio and video from community models, track predictions

  表示
• 

  Hugging Face AIコネクタ

  Access thousands of pre-trained AI models for NLP, vision, and audio tasks with the largest open-source machin

  表示

おすすめ

• 

  Hugging Face AIコネクタ

  Access thousands of pre-trained AI models for NLP, vision, and audio tasks with the largest open-source machin

  表示
• 

  Clarifai (Vision AI) AIコネクタ

  Manage AI inference via Clarifai — list apps, models, and workflows, and perform computer vision predictions d

  表示
• 

  Lambda Labs (GPU Cloud) AIコネクタ

  Manage AI infrastructure via Lambda Labs — launch GPU instances, monitor ML workloads, and manage SSH keys.

  表示
• 

  GPU Inference Memory Calculator AIコネクタ

  Estimate GPU VRAM requirements for LLM inference based on model parameters, precision, and batch size.

  表示