The POMA Grill MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your POMA Grill account. It provides structured and secure access to your project's indexed documents and ingestion tasks, so your agent can ingest content, search documents, manage metadata, monitor processing, and control bucket imports on your behalf.
- Document ingestion and tracking: Have your agent submit files or public web content for indexing, then monitor each ingestion job until processing is complete, blocked, or failed.
- Context search and retrieval: Instruct your agent to search across the project or within a specific document and return prompt-ready context or matching content chunks.
- Indexed document management: Let the agent list documents, inspect their metadata, update typed attributes, or permanently remove selected documents and their stored content.
- Cloud bucket ingestion: Direct your agent to preview and submit matching content from S3 or Google Cloud Storage buckets, then check progress, processing rates, credential status, and recent failures.
- Ingestion recovery and control: Have your agent pause, resume, or cancel bucket ingestion, update source credentials, review failed items, retry selected failures, or cancel document ingestion jobs.