This project is part of pgEdge Labs and is under active development. APIs and features may change without notice.
pgedge starfleet managed command reference
This page lists the usage, description and declared flags for every command in the managed sub-tree, which runs single-master databases pgEdge hosts and operates. There is no cluster or node to choose.
The managed tree shares the starfleet module's one connection: the flags on pgedge starfleet and the login it describes apply here unchanged. Global flags are declared on the root command and listed on the pgedge reference page.
For workflows and behavioral detail, run
pgedge llms starfleet managed against your installed binary.
pgedge starfleet managed
Usage: pgedge starfleet managed <command>
Manage pgEdge Managed databases
Example:
pgedge starfleet auth login
pgedge starfleet managed database list
pgedge starfleet managed backup
Usage: pgedge starfleet managed backup <command>
Inspect, take and restore pgEdge Managed backups
Example:
pgedge starfleet managed backup list --database-id <database_id>
pgedge starfleet managed backup create --database-id <database_id> --kind hot
pgedge starfleet managed backup restore <backup_id>
pgedge starfleet managed backup create
Usage: pgedge starfleet managed backup create [flags]
Take an on-demand backup
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--database-id string |
Yes | Database to back up (full UUID) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--kind string |
Yes | Backup tier to take: hot or durable |
Example:
pgedge starfleet managed backup create \
--database-id e5f6a7b8-c9d0-1234-efab-567890123456 --kind hot
pgedge starfleet managed backup create \
--database-id e5f6a7b8-c9d0-1234-efab-567890123456 --kind durable
pgedge starfleet managed backup get
Usage: pgedge starfleet managed backup get <backup_id>
Show backup details
Example:
pgedge starfleet managed backup get <backup_id>
pgedge starfleet managed backup get b8c9d0e1-f2a3-4567-bcde-678901234567 -o yaml
pgedge starfleet managed backup list
Usage: pgedge starfleet managed backup list [flags]
List backups
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--created-after string |
No | Only backups created at or after this RFC3339 time | |
--created-before string |
No | Only backups created at or before this RFC3339 time | |
--database-id string |
No | Filter by database (full UUID) | |
--descending |
No | true |
Sort newest first (--descending=false for oldest first) |
--kind string |
No | Filter by kind (e.g. hot, durable) | |
--limit int |
No | Maximum number of results to return (1-100, and 100 is also the default) | |
--offset int |
No | Offset into the results for pagination |
Example:
pgedge starfleet managed backup list
pgedge starfleet managed backup list \
--database-id e5f6a7b8-c9d0-1234-efab-567890123456 --kind durable
pgedge starfleet managed backup list --created-after 2026-08-01T00:00:00Z
pgedge starfleet managed backup restore
Usage: pgedge starfleet managed backup restore <backup_id> [flags]
Restore a database from a backup
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed backup restore <backup_id>
pgedge starfleet managed backup restore <backup_id> --force --wait
pgedge starfleet managed client-ip
Usage: pgedge starfleet managed client-ip
Show the source address the API sees for you
Example:
pgedge starfleet managed client-ip
pgedge starfleet managed database allowlist add <database_id> --my-ip
pgedge starfleet managed database
Usage: pgedge starfleet managed database <command>
Manage pgEdge Managed databases
Example:
pgedge starfleet managed database list
pgedge starfleet managed database get e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database allowlist
Usage: pgedge starfleet managed database allowlist <command>
Manage which addresses may reach a database
Example:
pgedge starfleet managed database allowlist get <database_id>
pgedge starfleet managed database allowlist add <database_id> --my-ip
pgedge starfleet managed database allowlist add <database_id> \
203.0.113.0/24 --label office --service mcp
pgedge starfleet managed database allowlist add
Usage: pgedge starfleet managed database allowlist add <database_id> [<cidr>...] [flags]
Allow addresses to reach an endpoint
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--label string |
No | Note stored on each rule added by this call, at most 64 characters | |
--my-ip |
No | Also allow the address the API sees this command arriving from | |
--service string |
No | Edit this service's own allowlist instead of the Postgres endpoint's: mcp, rag or postgrest | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database allowlist add <database_id> --my-ip
pgedge starfleet managed database allowlist add <database_id> \
203.0.113.0/24 198.51.100.9 --label office
pgedge starfleet managed database allowlist add <database_id> \
203.0.113.7 --service mcp --wait
pgedge starfleet managed database allowlist clear
Usage: pgedge starfleet managed database allowlist clear <database_id> [flags]
Close an endpoint to every address
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--service string |
No | Edit this service's own allowlist instead of the Postgres endpoint's: mcp, rag or postgrest | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database allowlist clear <database_id>
pgedge starfleet managed database allowlist clear <database_id> \
--service rag --force
pgedge starfleet managed database allowlist get
Usage: pgedge starfleet managed database allowlist get [<database_id>] [flags]
Show an endpoint's allowlist and its state
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--service string |
No | Edit this service's own allowlist instead of the Postgres endpoint's: mcp, rag or postgrest |
Example:
pgedge starfleet managed database allowlist get <database_id>
pgedge starfleet managed database allowlist get <database_id> --service mcp
pgedge starfleet managed database allowlist open
Usage: pgedge starfleet managed database allowlist open <database_id> [flags]
Let every address reach an endpoint
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--service string |
No | Edit this service's own allowlist instead of the Postgres endpoint's: mcp, rag or postgrest | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database allowlist open <database_id>
pgedge starfleet managed database allowlist open <database_id> --service mcp
pgedge starfleet managed database allowlist remove
Usage: pgedge starfleet managed database allowlist remove <database_id> <cidr>... [flags]
Stop addresses reaching an endpoint
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt when removing the last rule | |
--service string |
No | Edit this service's own allowlist instead of the Postgres endpoint's: mcp, rag or postgrest | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database allowlist remove <database_id> 203.0.113.7
pgedge starfleet managed database allowlist remove <database_id> \
198.51.100.0/24 --service rag
pgedge starfleet managed database allowlist set
Usage: pgedge starfleet managed database allowlist set <database_id> <cidr>... [flags]
Replace an endpoint's allowlist
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--label string |
No | Note stored on every rule set by this call, at most 64 characters | |
--service string |
No | Edit this service's own allowlist instead of the Postgres endpoint's: mcp, rag or postgrest | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database allowlist set <database_id> \
203.0.113.0/24 198.51.100.9 --label ci
pgedge starfleet managed database branch
Usage: pgedge starfleet managed database branch <command>
Manage branches of a managed database
Example:
pgedge starfleet managed database branch list <database_id>
pgedge starfleet managed database branch create <database_id> \
--display-name dev-copy
pgedge starfleet managed database branch get <database_id> <branch_id>
pgedge starfleet managed database branch metrics <database_id> <branch_id>
pgedge starfleet managed database branch logs <database_id> <branch_id>
pgedge starfleet managed database branch delete <database_id> <branch_id>
pgedge starfleet managed database branch create
Usage: pgedge starfleet managed database branch create <database_id> [flags]
Create a branch of a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--display-name string |
No | Display name for the branch, at most 25 characters | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database branch create e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database branch create e5f6a7b8-c9d0-1234-efab-567890123456 \
--display-name dev-copy
pgedge starfleet managed database branch delete
Usage: pgedge starfleet managed database branch delete <database_id> <branch_id> [flags]
Delete a branch
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database branch delete <database_id> <branch_id>
pgedge starfleet managed database branch delete <database_id> <branch_id> \
--force
pgedge starfleet managed database branch get
Usage: pgedge starfleet managed database branch get <database_id> <branch_id> [flags]
Show branch details
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--user-type string |
No | Role whose credentials to return: admin, app or app_read_only (default app) |
Example:
pgedge starfleet managed database branch get <database_id> <branch_id>
pgedge starfleet managed database branch get <database_id> <branch_id> \
--user-type admin -o yaml
pgedge starfleet managed database branch list
Usage: pgedge starfleet managed database branch list [<database_id>] [flags]
List a database's branches
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--descending |
No | List the newest branches first | |
--include-deleted |
No | Also show deleted branches | |
--limit int |
No | Maximum number of results to return (1-100, and 100 is also the default) | |
--offset int |
No | Offset into the results for pagination |
Example:
pgedge starfleet managed database branch list e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database branch list e5f6a7b8-c9d0-1234-efab-567890123456 \
--include-deleted
pgedge starfleet managed database branch logs
Usage: pgedge starfleet managed database branch logs <database_id> <branch_id> [flags]
Read a branch's logs
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--end-time string |
No | Window end as an RFC3339 timestamp | |
--max-lines int |
No | Maximum log records to return, 1 to 1000 (API default 100) | |
--start-time string |
No | Window start as an RFC3339 timestamp |
Example:
pgedge starfleet managed database branch logs e5f6a7b8-c9d0-1234-efab-567890123456 b1c2d3e4-f5a6-7890-bcde-f12345678901
pgedge starfleet managed database branch logs e5f6a7b8-c9d0-1234-efab-567890123456 b1c2d3e4-f5a6-7890-bcde-f12345678901 \
--max-lines 500
pgedge starfleet managed database branch metrics
Usage: pgedge starfleet managed database branch metrics <database_id> <branch_id> [flags]
Read a branch's metrics
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--end-time string |
No | Window end as an RFC3339 timestamp | |
--start-time string |
No | Window start as an RFC3339 timestamp | |
--window string |
No | Relative lookback window, in value,unit form (1 or more) |
Example:
pgedge starfleet managed database branch metrics e5f6a7b8-c9d0-1234-efab-567890123456 b1c2d3e4-f5a6-7890-bcde-f12345678901
pgedge starfleet managed database branch metrics e5f6a7b8-c9d0-1234-efab-567890123456 b1c2d3e4-f5a6-7890-bcde-f12345678901 \
--window 5,minutes
pgedge starfleet managed database connection-string
Usage: pgedge starfleet managed database connection-string [<database_id>] [flags]
Print a connection string for a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--format string |
No | uri |
Shape of the text output: uri or env |
--no-password |
No | Leave the password out of the string | |
--user-type string |
No | Role whose credentials to use: admin, app or app_read_only (default app) |
Example:
pgedge starfleet managed database connection-string e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database connection-string e5f6a7b8-c9d0-1234-efab-567890123456 \
--user-type admin --no-password
pgedge starfleet managed database connection-string e5f6a7b8-c9d0-1234-efab-567890123456 \
--format env > .env
pgedge starfleet managed database create
Usage: pgedge starfleet managed database create [flags]
Create a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--allow stringArray |
No | IPv4 address or CIDR block allowed to reach the Postgres endpoint; repeat for more than one | |
--display-name string |
No | Display name for the database, at most 25 characters | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--link |
No | Link the current folder to the new database (needs --wait or --follow) | |
--my-ip |
No | Also allow the address the API sees this command arriving from | |
--name string |
Yes | Database name | |
--open |
No | Admit every address (one 0.0.0.0/0 rule); never the default | |
--options strings |
No | Comma-separated list of options | |
--pg-version string |
No | Postgres major version to create the database on: 18, 17, 16 (fixed for the life of the database; defaults to the newest supported version) | |
--region string |
No | Region to provision the database in (e.g. us-east-1); checked against 'region list'; optional while the API publishes only one, and required as soon as it publishes more | |
--size string |
Yes | Managed size name (e.g. small, large); checked against 'size list' | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database create --name mydb \
--region us-east-1 --size small
pgedge starfleet managed database create --name mydb \
--region us-east-1 --size large --pg-version 16
pgedge starfleet managed database create --name mydb \
--region us-east-1 --size small --my-ip
pgedge starfleet managed database create --name myapp \
--size small --my-ip --wait --link
pgedge starfleet managed database delete
Usage: pgedge starfleet managed database delete <database_id> [flags]
Delete a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--delete-branches |
No | Also delete every branch first; data cannot be recovered | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database delete e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database delete e5f6a7b8-c9d0-1234-efab-567890123456 \
--force
pgedge starfleet managed database delete e5f6a7b8-c9d0-1234-efab-567890123456 \
--delete-branches --force
pgedge starfleet managed database env
Usage: pgedge starfleet managed database env <command>
Write a database's connection into a .env file
Example:
pgedge starfleet managed database env pull
pgedge starfleet managed database env pull
Usage: pgedge starfleet managed database env pull [<database_id>] [flags]
Write DATABASE_URL for a database into .env
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--branch string |
No | Pull this branch of the database instead of the linked target, leaving the link as it is | |
--file string |
No | File to write (default .env beside .pgedge/, or in the current folder when an ID is given) | |
--user-type string |
No | Role whose credentials to use: admin, app or app_read_only (default app) | |
--var string |
No | DATABASE_URL |
Variable name to set |
Example:
pgedge starfleet managed database env pull
pgedge starfleet managed database env pull e5f6a7b8-c9d0-1234-efab-567890123456 \
--file .env.local --user-type app_read_only
pgedge starfleet managed database env pull --branch 0a1b2c3d-4e5f-6789-abcd-ef0123456789 \
--file .env.preview
pgedge starfleet managed database get
Usage: pgedge starfleet managed database get [<database_id>] [flags]
Show managed database details
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--user-type string |
No | Role whose credentials to return: admin, app or app_read_only (default app) |
Example:
pgedge starfleet managed database get e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database get e5f6a7b8-c9d0-1234-efab-567890123456 \
--user-type admin -o yaml
pgedge starfleet managed database inspect
Usage: pgedge starfleet managed database inspect [<database_id>] <analysis> [flags]
Run a read-only diagnostic against a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--user-type string |
No | Role to connect as: admin, app or app_read_only (default app; admin for the statistics analyses) |
Example:
pgedge starfleet managed database inspect e5f6a7b8-c9d0-1234-efab-567890123456 table-sizes
pgedge starfleet managed database inspect e5f6a7b8-c9d0-1234-efab-567890123456 \
long-running-queries --user-type admin -o json
pgedge starfleet managed database link
Usage: pgedge starfleet managed database link [<database_id>] [flags]
Link the current folder to a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--branch string |
No | Link a branch of the database instead of the database itself | |
--force |
No | Replace a link to another database |
Example:
pgedge starfleet managed database link
pgedge starfleet managed database link e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database link e5f6a7b8-c9d0-1234-efab-567890123456 \
--branch 0a1b2c3d-4e5f-6789-abcd-ef0123456789
pgedge starfleet managed database list
Usage: pgedge starfleet managed database list [flags]
List managed databases
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--descending |
No | Sort in descending order | |
--limit int |
No | Maximum number of results to return (1-1000) | |
--offset int |
No | Offset into the results for pagination | |
--region string |
No | Filter by region |
Example:
pgedge starfleet managed database list
pgedge starfleet managed database list --region us-east-1 -o json
pgedge starfleet managed database logs
Usage: pgedge starfleet managed database logs [<database_id>] [flags]
Read a managed database's logs
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--end-time string |
No | Window end as an RFC3339 timestamp | |
--max-lines int |
No | Maximum log records to return, 1 to 1000 (API default 100) | |
--start-time string |
No | Window start as an RFC3339 timestamp |
Example:
pgedge starfleet managed database logs e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database logs e5f6a7b8-c9d0-1234-efab-567890123456 \
--max-lines 500
pgedge starfleet managed database logs e5f6a7b8-c9d0-1234-efab-567890123456 -o json
pgedge starfleet managed database logs e5f6a7b8-c9d0-1234-efab-567890123456 \
--start-time 2026-08-17T00:00:00Z
pgedge starfleet managed database mcp
Usage: pgedge starfleet managed database mcp <command>
Manage the MCP server on a database
Example:
pgedge starfleet managed database mcp deploy <database_id>
pgedge starfleet managed database mcp update <database_id> --allow-writes
pgedge starfleet managed database mcp deploy
Usage: pgedge starfleet managed database mcp deploy <database_id> [flags]
Deploy an MCP server on a database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--allow-writes |
No | Grant the MCP service read-write access (WARNING: allows LLM to modify data) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--embedding-api-key string |
No | API key for the embedding provider (required when --embedding-provider is set on deploy; the stored key is reused if omitted on update) | |
--embedding-model string |
No | Embedding model identifier (required when --embedding-provider is set) | |
--embedding-provider string |
No | Embedding provider: openai or voyage | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--init-tokens string |
No | Bearer token forwarded to the MCP server as INIT_TOKENS | |
--init-users string |
No | Comma-separated username:password pairs forwarded as INIT_USERS | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database mcp deploy <database_id>
pgedge starfleet managed database mcp deploy <database_id> \
--embedding-provider openai \
--embedding-model text-embedding-3-small \
--embedding-api-key sk-...
pgedge starfleet managed database mcp update
Usage: pgedge starfleet managed database mcp update <database_id> [flags]
Update the MCP server on a database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--allow-writes |
No | Grant the MCP service read-write access (WARNING: allows LLM to modify data) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--embedding-api-key string |
No | API key for the embedding provider (required when --embedding-provider is set on deploy; the stored key is reused if omitted on update) | |
--embedding-model string |
No | Embedding model identifier (required when --embedding-provider is set) | |
--embedding-provider string |
No | Embedding provider: openai or voyage | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--init-tokens string |
No | Bearer token forwarded to the MCP server as INIT_TOKENS | |
--init-users string |
No | Comma-separated username:password pairs forwarded as INIT_USERS | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database mcp update <database_id> --allow-writes
pgedge starfleet managed database mcp update <database_id> --allow-writes=false
pgedge starfleet managed database metrics
Usage: pgedge starfleet managed database metrics [<database_id>] [flags]
Read a managed database's metrics
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--end-time string |
No | Window end as an RFC3339 timestamp | |
--start-time string |
No | Window start as an RFC3339 timestamp | |
--window string |
No | Relative lookback window, in value,unit form (1 or more) |
Example:
pgedge starfleet managed database metrics e5f6a7b8-c9d0-1234-efab-567890123456
pgedge starfleet managed database metrics e5f6a7b8-c9d0-1234-efab-567890123456 \
--window 5,minutes
pgedge starfleet managed database metrics e5f6a7b8-c9d0-1234-efab-567890123456 -o json
pgedge starfleet managed database metrics e5f6a7b8-c9d0-1234-efab-567890123456 \
--start-time 2026-08-17T00:00:00Z
pgedge starfleet managed database postgrest
Usage: pgedge starfleet managed database postgrest <command>
Manage PostgREST on a database (not yet supported)
Example:
pgedge starfleet managed database postgrest deploy <database_id> \
--db-schemas public --db-anon-role app_read_only
pgedge starfleet managed database postgrest update <database_id> --max-rows 500
pgedge starfleet managed database postgrest deploy
Usage: pgedge starfleet managed database postgrest deploy <database_id> [flags]
Deploy PostgREST on a database (not yet supported)
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--cors-origins string |
No | Comma-separated list of allowed CORS origins | |
--db-anon-role string |
Yes | Postgres role used for unauthenticated requests | |
--db-pool int |
No | Database connections to keep open (1-30, default 10) | |
--db-schemas string |
Yes | Comma-separated schemas to expose as REST (e.g. public,api) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--jwt-audience string |
No | JWT audience claim to require | |
--jwt-role-claim-key string |
No | JSONPath to the role claim inside the JWT | |
--jwt-secret string |
No | JWT signing secret (min 32 chars); write-only, never returned | |
--max-rows int |
No | Maximum rows returned per request (1-10000, default 1000) | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database postgrest deploy <database_id> \
--db-schemas public --db-anon-role app_read_only
pgedge starfleet managed database postgrest deploy <database_id> \
--db-schemas public,api --db-anon-role app_read_only \
--jwt-secret <32-char-minimum-secret>
pgedge starfleet managed database postgrest update
Usage: pgedge starfleet managed database postgrest update <database_id> [flags]
Update PostgREST on a database (not yet supported)
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--cors-origins string |
No | Comma-separated list of allowed CORS origins | |
--db-anon-role string |
No | Postgres role used for unauthenticated requests | |
--db-pool int |
No | Database connections to keep open (1-30, default 10) | |
--db-schemas string |
No | Comma-separated schemas to expose as REST (e.g. public,api) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--jwt-audience string |
No | JWT audience claim to require | |
--jwt-role-claim-key string |
No | JSONPath to the role claim inside the JWT | |
--jwt-secret string |
No | JWT signing secret (min 32 chars); write-only, never returned | |
--max-rows int |
No | Maximum rows returned per request (1-10000, default 1000) | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database postgrest update <database_id> --max-rows 500
pgedge starfleet managed database postgrest update <database_id> \
--db-schemas public,api
pgedge starfleet managed database rag
Usage: pgedge starfleet managed database rag <command>
Manage the RAG server on a database
Example:
pgedge starfleet managed database rag deploy <database_id> \
--embedding-llm-provider openai \
--embedding-llm-model text-embedding-3-small \
--embedding-llm-api-key "$OPENAI_API_KEY" \
--completion-llm-provider openai \
--completion-llm-model gpt-4o \
--completion-llm-api-key "$OPENAI_API_KEY" \
--pipeline-config pipelines.json
pgedge starfleet managed database rag update <database_id> --top-n 10
pgedge starfleet managed database rag deploy
Usage: pgedge starfleet managed database rag deploy <database_id> [flags]
Deploy a RAG server on a database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--completion-llm-api-key string |
Yes | API key for the completion LLM provider | |
--completion-llm-model string |
Yes | Completion LLM model identifier | |
--completion-llm-provider string |
Yes | Completion LLM provider (e.g. openai, anthropic) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--embedding-llm-api-key string |
Yes | API key for the embedding LLM provider | |
--embedding-llm-model string |
Yes | Embedding LLM model identifier | |
--embedding-llm-provider string |
Yes | Embedding LLM provider (openai or anthropic) | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--pipeline-config string |
Yes | Path to a JSON file containing pipeline definitions | |
--token-budget int |
No | Default max completion tokens across all pipelines | |
--top-n int |
No | Default number of results to retrieve per pipeline | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database rag deploy <database_id> \
--embedding-llm-provider openai \
--embedding-llm-model text-embedding-3-small \
--embedding-llm-api-key sk-... \
--completion-llm-provider openai --completion-llm-model gpt-4o \
--completion-llm-api-key sk-... --pipeline-config pipelines.json
pgedge starfleet managed database rag update
Usage: pgedge starfleet managed database rag update <database_id> [flags]
Update the RAG server on a database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--completion-llm-api-key string |
No | API key for the completion LLM provider | |
--completion-llm-model string |
No | Completion LLM model identifier | |
--completion-llm-provider string |
No | Completion LLM provider (e.g. openai, anthropic) | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--embedding-llm-api-key string |
No | API key for the embedding LLM provider | |
--embedding-llm-model string |
No | Embedding LLM model identifier | |
--embedding-llm-provider string |
No | Embedding LLM provider (openai or anthropic) | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--pipeline-config string |
No | Path to a JSON file containing pipeline definitions | |
--token-budget int |
No | Default max completion tokens across all pipelines | |
--top-n int |
No | Default number of results to retrieve per pipeline | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database rag update <database_id> --top-n 10
pgedge starfleet managed database rag update <database_id> \
--pipeline-config pipelines.json
pgedge starfleet managed database resize
Usage: pgedge starfleet managed database resize <database_id> [flags]
Resize a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--size string |
Yes | Target managed size name (e.g. small, large); checked against 'size list' | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database resize e5f6a7b8-c9d0-1234-efab-567890123456 \
--size large
pgedge starfleet managed database resize e5f6a7b8-c9d0-1234-efab-567890123456 \
--size large --force --wait
pgedge starfleet managed database rotate-password
Usage: pgedge starfleet managed database rotate-password <database_id> [flags]
Rotate a built-in role's password
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--role string |
Yes | Built-in role to rotate: admin, app, or app_read_only | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database rotate-password e5f6a7b8-c9d0-1234-efab-567890123456 --role app
pgedge starfleet managed database rotate-password e5f6a7b8-c9d0-1234-efab-567890123456 --role admin --force
pgedge starfleet managed database service
Usage: pgedge starfleet managed database service <command>
Manage services deployed on a database
Example:
pgedge starfleet managed database service list <database_id>
pgedge starfleet managed database service get <database_id> mcp
pgedge starfleet managed database service get
Usage: pgedge starfleet managed database service get <database_id> <type>
Show details of a service
Example:
pgedge starfleet managed database service get <database_id> mcp
pgedge starfleet managed database service get <database_id> rag -o json
pgedge starfleet managed database service list
Usage: pgedge starfleet managed database service list <database_id>
List services deployed on a database
Example:
pgedge starfleet managed database service list <database_id>
pgedge starfleet managed database service list <database_id> -o json
pgedge starfleet managed database service remove
Usage: pgedge starfleet managed database service remove <database_id> <type> [flags]
Remove a service type from a database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--follow |
No | Stream the task's step messages until it reaches a terminal state | |
--force |
No | Skip the confirmation prompt | |
--wait |
No | Wait for the operation's task to reach a terminal state | |
--wait-interval int |
No | 5 |
Polling interval in seconds when --wait/--follow is set |
--wait-timeout int |
No | 600 |
Max seconds to wait when --wait/--follow is set |
Example:
pgedge starfleet managed database service remove <database_id> mcp
pgedge starfleet managed database service remove <database_id> rag --force
pgedge starfleet managed database unlink
Usage: pgedge starfleet managed database unlink
Remove the current folder's database link
Example:
pgedge starfleet managed database unlink
pgedge starfleet managed database update
Usage: pgedge starfleet managed database update <database_id> [flags]
Update a managed database
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--deletion-protection |
No | Refuse deletion until this is turned off again | |
--display-name string |
No | Display name for the database, at most 25 characters | |
--dry-run string[="checks"] |
No | Run every client-side check, then stop before sending the write and report the request that would have been sent. Checks that read the API do run, so this needs credentials; the server validates nothing until the real write | |
--options strings |
No | Comma-separated list of options |
Example:
pgedge starfleet managed database update e5f6a7b8-c9d0-1234-efab-567890123456 \
--display-name "My Database"
pgedge starfleet managed database update e5f6a7b8-c9d0-1234-efab-567890123456 \
--options key1,key2
pgedge starfleet managed database update e5f6a7b8-c9d0-1234-efab-567890123456 \
--deletion-protection
pgedge starfleet managed database update e5f6a7b8-c9d0-1234-efab-567890123456 \
--deletion-protection=false
pgedge starfleet managed pg-version
Usage: pgedge starfleet managed pg-version <command>
Discover Postgres versions for managed databases
Example:
pgedge starfleet managed pg-version list
pgedge starfleet managed pg-version list
Usage: pgedge starfleet managed pg-version list
List available Postgres versions
Example:
pgedge starfleet managed pg-version list
pgedge starfleet managed pg-version list -o json
pgedge starfleet managed region
Usage: pgedge starfleet managed region <command>
Discover pgEdge Managed regions
Example:
pgedge starfleet managed region list
pgedge starfleet managed region list
Usage: pgedge starfleet managed region list
List available regions
Example:
pgedge starfleet managed region list
pgedge starfleet managed region list -o json
pgedge starfleet managed size
Usage: pgedge starfleet managed size <command>
Discover pgEdge Managed sizes
Example:
pgedge starfleet managed size list
pgedge starfleet managed size list --pricing
pgedge starfleet managed size get
Usage: pgedge starfleet managed size get <size_id>
Show a size's details
Example:
pgedge starfleet managed size get <size_id>
pgedge starfleet managed size get <size_id> -o yaml
pgedge starfleet managed size list
Usage: pgedge starfleet managed size list [flags]
List available sizes
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--pricing |
No | Include each size's current price |
Example:
pgedge starfleet managed size list
pgedge starfleet managed size list --pricing
pgedge starfleet managed size list -o json
pgedge starfleet managed task
Usage: pgedge starfleet managed task <command>
Inspect pgEdge Managed tasks
Example:
pgedge starfleet managed task list --subject-id <database_id>
pgedge starfleet managed task wait <task_id>
pgedge starfleet managed task get
Usage: pgedge starfleet managed task get <task_id>
Show task details
Example:
pgedge starfleet managed task get <task_id> -o yaml
pgedge starfleet managed task list
Usage: pgedge starfleet managed task list [flags]
List tasks
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--limit int |
No | Maximum number of results to return (API default 25) | |
--name string |
No | Filter by task name (e.g. rotate-password-managed) | |
--offset int |
No | Offset into the results for pagination | |
--status string |
No | Filter by status (queued, running, succeeded, failed) | |
--subject-id string |
No | Filter by subject (full UUID) | |
--subject-kind string |
No | Filter by subject kind |
Example:
pgedge starfleet managed task list
pgedge starfleet managed task list --subject-id <database_id> --status failed
pgedge starfleet managed task wait
Usage: pgedge starfleet managed task wait <task_id> [flags]
Wait for a task to complete
Flags:
| Flag | Required | Default | Description |
|---|---|---|---|
--follow |
No | Stream the task's step messages instead of bare status lines | |
--wait-interval int |
No | 5 |
Polling interval in seconds |
--wait-timeout int |
No | 600 |
Maximum seconds to wait for task completion |
Example:
pgedge starfleet managed task wait <task_id> --wait-timeout 900
pgedge starfleet managed task wait <task_id> --follow