CX - Coralogix CLI
The observability backbone for AI agents and engineering teams.
Connect your agents to live logs, traces, metrics, dashboards, and alerts so they can investigate incidents, explain what changed, and reason about production with real operational context.
What you can do
- Query any signal-logs, metrics, spans, and RUM data-with DataPrime or PromQL, and render results as tables, raw JSON, or a token-efficient format for AI agents.
- Manage the full Coralogix stack: alerts, incidents, notifications, IAM, SLOs, dashboards, data pipeline rules, TCO policies, and more.
- Run the same command across multiple profiles or regions in a single invocation with multi-profile fan-out.
- Give your AI agent a single entry point to production observability:
cx schemadumps the entire command tree as JSON so agents can self-discover capabilities without manual documentation. - Find the right log or span field by describing it in natural language.
- Browse the DataPrime language reference offline.
- Plug Coralogix into your AI coding agent with bundled skills for Claude Code, Cursor, Codex, and 40+ more agents.
Features
- DataPrime and PromQL at the terminal-Coralogix's proprietary query languages work end-to-end without leaving the shell.
- 27 commands across 9 domains-from querying signals to managing IAM, notifications, TCO, and archiving-all in one binary.
- Multi-profile fan-out with
-p prod-eu -p prod-us <command>-run one command across multiple accounts or regions in a single invocation, with rows tagged by profile. agentsoutput format-token-efficient JSON that auto-spills to a temp file once the serialized payload exceeds 100 KiB, so AI agents get a path instead of a flooded context window.cx schema-outputs the full command tree as structured JSON, purpose-built for agent discovery with no help-text parsing required.- Semantic field search-find the right log or span field by describing it in natural language.
- Bundled skills for Claude Code, Cursor, Codex, OpenCode, and 40+ more agents, distributed via
npx skills add.
Install cx for AI agents
Most macOS users should use Homebrew. If you are not on macOS, use the cross-platform install script.
1. Install the CLI
Recommended for macOS: Homebrew
2. Install agent skills
Skills teach your AI agent how to use cx for observability investigations.
3. Optional: install shell autocomplete
Most macOS users use zsh:
For other shells, see Shell completions. For Cargo, pre-built binaries, Nix, and source builds, see Installation reference.
Quick start
Follow these steps to go from a fresh install to a working query.
Create a profile.
cx profiles addopens an interactive prompt for region, credentials, and default output format:Query logs. The positional argument is a DataPrime query:
Query metrics.
cx metrics querytakes a PromQL expression:Search distributed spans. The positional argument is a DataPrime filter;
source spansis prepended automatically:List dashboards to confirm the API is reachable:
Try semantic search to find dashboards or queries:
Run Commands are grouped by domain. Run Query Observe Detect & Respond Notifications Data Pipeline Cost & Storage Integrations Access Agent & Localcx <command> --help for full syntax and examples on any command.Commands
cx --help for the full organized listing, or cx schema for a machine-readable JSON tree.Command Purpose cx logsQuery logs using DataPrime cx spansQuery distributed spans cx metricsQuery metrics using PromQL: query, query-range, search, get-labelscx dataprimeDataPrime language reference and raw queries: list, show, querycx search-fieldsFind log or span fields by description (default) or by value content ( -s value)Command Purpose cx dashboardsManage or search dashboards and folders cx viewsManage saved views and view folders cx slosManage SLO definitions Command Purpose cx alertsManage alerts: list, get, create, enable, disable, events, event-stats, suppression-rulescx incidentsManage and triage incidents Command Purpose cx notificationsManage connectors, routers, presets, and notification testing cx webhooksManage outgoing webhooks and automation actions Command Purpose cx parsing-rulesManage log parsing rules cx enrichmentsManage enrichment rules and custom enrichment tables cx e2mManage Events2Metrics definitions cx recording-rulesManage Prometheus recording rule groups Command Purpose cx usageView data usage and consumption metrics cx tcoManage TCO policies and settings cx retentionsManage data retention settings cx archive (risky)Manage data archive storage configuration Command Purpose cx integrationsManage integrations, extensions, and contextual data Command Purpose cx iam (risky)Manage API keys, roles, scopes, users, groups, and IP access Command Purpose cx schemaOutput the full command tree as JSON for agent consumption cx profilesManage profiles: list, add, delete, set-defaultcx completionsShell tab-completion: install, refresh, generatecx cleanupRemove cx_results* temp files older than 30 minutesGlobal options
Configuration
Configuration lives in ~/.cx/:
Credentials (API keys or OAuth tokens) are stored either inline in the profile TOML with 0600 permissions (credential_storage = "file", the default) or in the OS keyring (credential_storage = "os_store" - macOS Keychain, Windows Credential Manager, or D-Bus Secret Service on Linux). cx profiles add prompts for the choice. On Linux, keyring support requires a glibc build; the default install script and release binaries use musl, which has no keyring backend, so os_store is unavailable there.
Environment variables override profile settings: CX_PROFILE, CX_API_KEY, and CX_REGION.
See docs/configuration.md for the full reference.
Output formats
Choose an output format with -o or by setting the profile default.
text-human-readable tables with color. Default.json-raw, pretty-printed API responses for scripting.agents-token-efficient format for AI agents. Large responses automatically spill to a temporary file and the path is returned.
See docs/agents-output.md for the agents format specification.
AI agent skills
cx ships a companion skill bundle for Claude Code, Cursor, Codex, OpenCode, and 40+ other agents. The skills teach your agent how to investigate issues by querying Coralogix-without memorizing DataPrime syntax or API endpoints.
The install flow above installs all skills. Use these variants if you want to customize where or what you install.
Install selected skills:
Install globally for all projects:
Available skills: cx-query-logs, cx-query-spans, cx-metrics-query, cx-alerts, cx-dataprime, cx-rum, cx-telemetry-querying, cx-create-dashboard, cx-cost-optimization, cx-incident-management, cx-data-pipeline, cx-platform-admin, cx-observability-setup. See skills/README.md for per-skill usage.
Multi-profile fan-out
Repeat -p to run a command across multiple profiles in parallel. Results are merged and tagged with the profile name:
See docs/multi-profile.md for more examples.
Installation reference
Pin a CLI version
CX_VERSION=0.1.0 curl -fsSL https://raw.githubusercontent.com/coralogix/cx-cli/master/install.sh | sh
Cargo
Pre-built binaries
Download the latest release for your platform from GitHub Releases. The flake exposes both the Symlink each skill into Nix
nix run github:coralogix/cx-cli -- --help # try without installing
nix profile install github:coralogix/cx-cli # install into your profile
cx binary and the agent skill bundle:{
inputs.cx-cli.url = "github:coralogix/cx-cli";
outputs = { self, nixpkgs, cx-cli, ... }: {
# cx-cli.packages.${system}.default -> the `cx` binary
# cx-cli.packages.${system}.skills -> store path with all cx-* skills
};
}
Home Manager
~/.claude/skills/ (adjust the target path for other agents):## home.nix
{ inputs, pkgs, lib, ... }:
let
skills = inputs.cx-cli.packages.${pkgs.system}.skills;
in {
home.packages = [ inputs.cx-cli.packages.${pkgs.system}.default ];
home.file = lib.mapAttrs'
(name: _: lib.nameValuePair ".claude/skills/${name}" { source = "${skills}/${name}"; })
(lib.filterAttrs (_: t: t == "directory") (builtins.readDir skills));
}
Shell completions
cx supports tab-completion for all commands, flags, subcommands, and profile names.
Managed install (recommended)
Let cx install and track a completion script for you. It writes to a standard user-writable location and records the path so cx completions refresh can update it later:
Default paths used by each shell:
| Shell | Default path |
|---|---|
| zsh | ~/.zfunc/_cx |
| bash | ~/.local/share/bash-completion/completions/cx |
| fish | ~/.config/fish/completions/cx.fish |
After installing for zsh, add ~/.zfunc to your $fpath if it isn't already there (the install command will tell you):
Refreshing after profile changes
When you add or delete a profile, cx will remind you to refresh if you have managed completions. You can run it any time:
Only files previously installed by cx completions install are updated.
Manual install
To generate a script yourself and pipe it anywhere:
cx completions generate zsh > ~/.zfunc/_cx
cx completions generate bash > ~/.local/share/bash-completion/completions/cx
cx completions generate fish > ~/.config/fish/completions/cx.fish
Dynamic completions (always-fresh profile names)
For profile names to update automatically on every Tab press without running refresh, source completions dynamically on each shell start. This calls back into cx at completion time:
zsh - add to ~/.zshrc:
bash - add to ~/.bashrc:
fish - add to ~/.config/fish/config.fish:
Migrating from cxctl
cx replaces the older Scala-based cxctl. If you are looking for documentation on the legacy tool, see the Coralogix CLI (legacy) docs. cx does not currently cover all legacy surfaces, including LiveTail and account invite flows.
Contributing
We welcome contributions! See CONTRIBUTING.md for the ownership model, PR review process, and step-by-step guides for adding commands and skills.
Further reading
License
Apache-2.0

