Infrastructure AI agents
can provision themselves

Deploy

# Upload your MCP server
curl -X POST https://api.eustore.dev/v1/mcp/deploy \
  -H "Authorization: Bearer $TOKEN" \
  -F "name=my-server" \
  -F "runtime=node" \
  -F "file=@server.tar.gz"

# Response:
{"endpoint": "https://mcp.eustore.dev/you/my-server"}

Connect

// Claude Desktop config
{
  "mcpServers": {
    "my-server": {
      "url": "https://mcp.eustore.dev/you/my-server/sse"
    }
  }
}

// Works with: Claude, Cursor, Windsurf, Cline

Create Bucket

curl -X POST https://api.eustore.dev/v1/storage/buckets \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"name":"my-data","region":"eu-central-fsn1"}'

# Returns: S3 endpoint + credentials
# First bucket is free

Upload with boto3

import boto3

s3 = boto3.client("s3",
    endpoint_url=creds["endpoint"],
    aws_access_key_id=creds["access_key"],
    aws_secret_access_key=creds["secret_key"])

s3.put_object(
    Bucket=creds["bucket"],
    Key="model.bin",
    Body=data)

Create + Insert

# Create collection
curl -X POST .../v1/vectors/collections \
  -H "Authorization: Bearer $TOKEN" \
  -d '{"name":"memory","vector_size":1536,
       "distance":"cosine"}'

# Insert vectors
curl -X POST .../v1/vectors/collections/memory/insert \
  -d '{"vectors":[[0.1,0.2,...]],
       "payloads":[{"text":"User prefers dark mode"}]}'

Search

# Semantic search
curl -X POST .../v1/vectors/collections/memory/search \
  -H "Authorization: Bearer $TOKEN" \
  -d '{"vector":[0.1,0.2,...],
       "limit":5}'

# Returns:
{"results": [
  {"score": 0.94,
   "payload": {"text": "User prefers dark mode"}}
]}