STRG MCP Server
Connect compatible agents to your STRG account through OAuth. You choose the access key, buckets and permissions the agent can use.
Review capabilities AI-NATIVE STORAGE
Object storage for AI agents
Connect external agents through the STRG MCP Server or use the STRG Assistant inside the console to design, configure and manage storage.
STRG AssistantAccount context enabled
Set up storage for a RAG application using private legal documents.
I’ll prepare the storage plan. Private EU-Central bucket Versioning and lifecycle policy Read/write key for ingestion Read-only key for retrieval
Ready for review
CONNECT FROMClaudeCodexCursorOther MCP clients
Two ways to work
Connect an agent or use the assistant
Use the same storage platform through a familiar S3 API, an MCP client or the assistant built into STRG.
STRG Assistant
Describe what you are building. The assistant can propose an architecture, explain the decisions and carry out approved account operations.
See what it can doSTRG MCP Server
Manage storage through MCP
The MCP server combines STRG documentation with permission-controlled access to your account. Agents can reason about the setup and perform the storage work you authorize.
Your permissions remain the boundaryOAuth, scoped keys and revocable access keep each agent within the access you approve.
BucketsCreate, list and inspect buckets
ObjectsUpload, retrieve, copy and organize objects
AccessCreate permission-scoped credentials
ConfigurationManage lifecycle and retention rules
LinksGenerate temporary upload and download URLs
KnowledgeUse STRG documentation while working
STRG Assistant
Create buckets and access keys in chat
The assistant is grounded in STRG documentation and your account configuration. It can guide the decision or carry out the approved steps for you.
Recommend bucket and prefix structures Plan RAG, agent-memory and data-pipeline storage Create buckets in the right region Configure Object Lock and lifecycle rules Create permission-scoped access keys Generate SDK and CLI configuration Diagnose integration problems Explain expected cost before changes are applied
01 / DESCRIBETell STRG what you are building
“I need private uploads for customer documents and a RAG pipeline.”
02 / REVIEWApprove the proposed architecture
Region, buckets, permissions, retention and estimated cost.
03 / CREATELet the assistant configure it
Buckets, access keys and integration instructions are prepared in the console.
Built for agent workloads
Storage patterns AI teams already need
Use standard object storage without building another control layer for your agents.
Long-term memory for agents
Keep notes, transcripts and files available between agent sessions.
Documents for RAG
Store the source documents your retrieval pipeline indexes and cites.
Models and datasets
Keep training data, checkpoints and evaluation sets in one object layer.
Agent-generated files
Give agents a controlled place to create, retrieve and share outputs.
Data pipelines
Separate raw, processed and published data with clear access boundaries.
Logs and event archives
Retain application and agent activity for analysis and review.
Access control
Choose what each agent can access
Agents act through credentials and permissions selected by the customer. Access can be scoped by bucket and operation, reviewed, and revoked.
OAuth authenticationScoped access keysCustomer-approved permissionsRevocable agent access
Connect an agent to STRG
Use MCP for external agents or build the storage setup with the STRG Assistant.