A command-line client for the Redash REST API — single static binary, works against any Redash instance (cloud or self-hosted).

Built in Go with cobra + viper. Outputs pretty tables by default, with --format json and --format csv for scripting.

• Ad-hoc queries — send SQL to a data source and print results (table, JSON, or CSV).
• Saved queries — list, get, run, create, update, archive.
• Parameters — pass query parameters via repeatable --param key=value flags or a single --params '{...}' JSON object.
• Data sources — list, get, fetch schema.
• Dashboards — list, get.
• Users — list, get, create, disable, enable.
• Multiple Redash instances — config file with named profiles, switchable via --profile or REDASH_PROFILE.
• Safe by default — --yes required for destructive operations; API keys redacted in config show.

brew install leroy/tap/redash

curl -sSfL https://raw.githubusercontent.com/leroy/redash/main/install.sh | sh

Honours VERSION (e.g. VERSION=v1.0.0) and INSTALL_DIR (default /usr/local/bin; uses sudo if needed):

curl -sSfL https://raw.githubusercontent.com/leroy/redash/main/install.sh \
  | VERSION=v1.1.0 INSTALL_DIR=$HOME/.local/bin sh

The script detects OS/arch, verifies the SHA-256 against checksums.txt, and refuses to install on mismatch.

go install github.com/leroy/redash@latest

Download from the Releases page (Linux, macOS, Windows — amd64 and arm64).

git clone https://github.com/leroy/redash
cd redash
go build -o redash .

# 1. Create a profile (interactive; prompts for URL + API key).
redash config init --name prod --default

# 2. Verify.
redash config show

# 3. List data sources to find the one you want to query.
redash datasources list

# 4. Run an ad-hoc query.
redash query --datasource 3 "SELECT count(*) FROM users"

# 5. Pipe SQL from a file or stdin.
cat report.sql | redash query -d 3 -f -
redash query -d 3 -f report.sql -o csv > report.csv

The config file lives at $XDG_CONFIG_HOME/redash-cli/config.yaml (falling back to ~/.config/redash-cli/config.yaml).

default_profile: prod
profiles:
  prod:
    url: https://redash.example.com
    api_key: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
    timeout: 30s
  staging:
    url: https://staging.redash.example.com
    api_key: yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy
    insecure: true

Override any field with environment variables:

The API key is a user-level key or a per-query key from the Redash UI. For admin-only endpoints (user management, creating data sources, etc.) you'll need an admin user's key.

Every command accepts --format table|json|csv (default table), --profile NAME, --config PATH, --timeout DURATION, and --insecure.

redash query --datasource 3 "SELECT 1"
redash query -d 3 -f query.sql --param country=US --param limit=100
redash query -d 3 "SELECT * FROM orders WHERE country = '{{country}}'" \
  --params '{"country":"US"}' \
  --max-age 300 \
  --format csv > orders.csv

• --datasource, -d — data source ID (required)
• --file, -f — read SQL from file (or - for stdin)
• --param k=v — repeatable query parameter (values auto-coerced to int/float/bool/string)
• --params '{...}' — full JSON parameters object (takes precedence over --param)
• --max-age SECS — accept a cached result this old (0 = always execute)
• --poll DURATION — job polling interval (default 500ms)

redash queries list --search "active users" --tag revenue
redash queries get 42
redash queries run 42 --param country=US
redash queries create --name "MRR" --datasource 3 -f mrr.sql --tag revenue
redash queries update 42 --name "MRR (v2)" --publish
redash queries archive 42 --yes

redash datasources list
redash datasources get 3
redash datasources schema 3
redash datasources schema 3 --refresh --format json

redash dashboards list --search revenue
redash dashboards get my-dashboard-slug

redash users list
redash users get 7
redash users create --name "Alice" --email alice@example.com --group 1
redash users disable 7
redash users enable 7

redash config path
redash config show
redash config init --name staging --url https://... --api-key ...
redash config set prod timeout 60s
redash config use staging
redash config remove old-profile

redash is designed to be driven by AI agents as a drop-in replacement for a Redash MCP server. The authoritative usage guide is compiled into the binary and is always version-matched:

redash manual                      # full agent-oriented markdown
redash manual --format json        # structured catalog for machine parsing
redash manual --topic query        # single section
redash manual --list-topics        # enumerate available topics

Run redash manual --format json once at startup — it's equivalent to an MCP server's list_tools + describe_tool handshake.

A short AGENTS.md at the repo root is auto-discovered by Cursor, Claude Code, and other coding agents; it points to the same command.

Drift is prevented by a unit test: every subcommand must have a manual entry, or CI fails.

• table (default) — ANSI-styled table with header, good for terminals.
• json — for query results, an array of row objects preserving native types (numbers as numbers, bools as bools). For get commands, the full raw Redash response.
• csv — RFC 4180 CSV with header row. Use for > file.csv pipelines.

• 0 — success
• 1 — any error (config, API, network, user error). Error message is written to stderr.

# Run the test suite.
go test -race ./...

# Vet.
go vet ./...

# Build a dev binary.
go build -o redash .
./redash --help

# Try a release build locally (requires goreleaser).
goreleaser release --snapshot --clean

.
├── main.go               # thin entry point
├── cmd/                  # cobra commands
│   ├── root.go
│   ├── query.go          # ad-hoc query
│   ├── queries.go        # saved queries CRUD + execute
│   ├── datasources.go
│   ├── dashboards.go
│   ├── users.go
│   ├── config.go
│   └── version.go
└── internal/
    ├── client/           # Redash HTTP client
    ├── config/           # config file + env overrides
    └── output/           # table/json/csv formatters

MIT — see LICENSE.