Manage Redis Cloud, Redis Enterprise, and Redis databases from one tool -- as a CLI for humans or an MCP server for AI agents.
redisctl replaces curl-and-jq scripts against the Redis Cloud and Enterprise REST APIs with a single binary that handles authentication, async polling, output formatting, and error handling. The same tool ships as an MCP server so AI assistants can manage Redis infrastructure directly.
| I want to... | Start here |
|---|---|
| Manage Redis Cloud from my terminal | CLI Quick Start |
| Manage Redis Enterprise from my terminal | CLI Quick Start |
| Let an AI assistant manage Redis | MCP Server |
| Query or inspect a Redis database directly | Database Tools |
# Homebrew (macOS/Linux)
brew install redis-developer/homebrew-tap/redisctl
# Cargo
cargo install redisctl
# Binary releases: https://github.com/redis-developer/redisctl/releases# Redis Cloud
redisctl profile set prod \
--deployment cloud \
--api-key "$REDIS_CLOUD_API_KEY" \
--api-secret "$REDIS_CLOUD_SECRET_KEY"
# Redis Enterprise
redisctl profile set dev \
--deployment enterprise \
--url "https://cluster.local:9443" \
--username "admin@redis.local" \
--password "$REDIS_ENTERPRISE_PASSWORD"# List databases (platform inferred from profile)
redisctl database list
# Create a database and wait for it to be ready
redisctl database create @db-config.json --wait
# Get cluster info as a table
redisctl cluster get -o table
# Filter output with JMESPath
redisctl database list -q 'databases[?status==`active`].name'The cloud/enterprise prefix is optional -- the CLI infers the platform from your profile. Use explicit prefixes (redisctl cloud database list) in scripts or when you have profiles for both platforms.
redisctl-mcp exposes 300+ tools to AI assistants (Claude Desktop, Cursor, VS Code, or any MCP client). It covers the full Cloud and Enterprise APIs plus direct Redis database operations -- all with a safety-first policy system.
{
"mcpServers": {
"redisctl": {
"command": "redisctl-mcp",
"args": ["--profile", "my-profile"]
}
}
}Read-only by default. Write and destructive operations require explicit opt-in via a policy file (redisctl-mcp.toml):
tier = "standard" # read-only (default) | standard (read-write) | full (destructive)REDIS_ENTERPRISE_PASSWORD is passed without a value so Docker forwards it from your host environment (keeps the secret out of the config file):
{
"mcpServers": {
"redisctl": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "REDIS_ENTERPRISE_URL=https://cluster:9443",
"-e", "REDIS_ENTERPRISE_USER=admin@redis.local",
"-e", "REDIS_ENTERPRISE_PASSWORD",
"ghcr.io/redis-developer/redisctl",
"redisctl-mcp"
]
}
}
}See the MCP documentation for client configuration guides, the full tool catalog, and safety policy reference.
Subscriptions, databases, VPC peering, Transit Gateway, PrivateLink, Private Service Connect, ACLs, cloud accounts, tasks, and async operations.
Clusters, nodes, shards, databases (BDBs), Active-Active (CRDBs), users, roles, LDAP, logs, metrics, alerts, support packages, and diagnostics.
Connect directly to any Redis instance for key inspection, data structure operations, server diagnostics, and health checks. Available in both the CLI and MCP server.
- Async operations --
--waitautomatically polls long-running operations to completion - Output formats -- tables, JSON, YAML, with JMESPath filtering (
-q) - Profiles -- manage multiple environments with optional keyring-backed credential storage
- Workflows -- high-level commands that compose multi-step operations (e.g.,
subscription-setup) - Raw API access --
redisctl api cloud get /subscriptions/12345for any endpoint - Streaming --
--followfor real-time log tailing
API client libraries (separate repositories):
git clone https://github.com/redis-developer/redisctl.git
cd redisctl
cargo build --release
cargo test --workspace
cargo clippy --all-targets -- -D warningsSee the Contributing Guide.
Licensed under either of Apache License 2.0 or MIT, at your option.