Nebius AI MCP Server

Nebius AI

Nebius AI Studio API for managing data files, batch processing jobs, and AI model discovery. Base URL: https://api.studio.nebius.com.

Entity hierarchy

• Files — uploaded data files in your repository. Each file has an ID, name, and size. Files are the input to batch processing jobs.
• File contents — the actual data rows/records within a file. Accessed by file_id. Use after listing files to inspect what's inside a specific file.
• Batches — data processing jobs. Each batch has an ID and status tracking its progress. Batches consume files as input.
• Models — available AI models with IDs, names, and capability descriptions. Browse to choose the right model for a task.

Workflows

• Data inspection: list files to see what's in your repository → pick a file → retrieve its contents to examine the data before processing.
• Batch monitoring: list batches to review processing job statuses and track which jobs are in progress, completed, or failed.
• Model selection: list available models to compare capabilities before deciding which model to apply to a batch processing task.

Gotchas

• file_id required for contents: you must list files first to get file IDs, then use the ID to retrieve file contents.
• No parameters on list endpoints: batches, files, and models list endpoints take no filter or pagination parameters — they return all items.
• Read-only: the current toolset supports listing and inspection only. Creating files, submitting batches, or running model inference requires additional API endpoints not yet exposed.