ff-cli
Standalone command-line tool for building, deploying, and operating FireFoundry agent bundles. Ships as a single binary with native Kubernetes and container integration — no runtime dependencies.
Why ff-cli
ff-cli is your primary interface for the FireFoundry platform. It ships as a single binary with zero dependencies. Unlike tools that require separate installations of kubectl, helm, or docker, ff-cli integrates Kubernetes, Helm, and container operations natively — giving you streamlined operations with proper error handling and richer output.
637+ unit tests ensure reliability across all packages. The clean three-layer architecture — CLI commands, business logic managers, and core executors — keeps the codebase maintainable while supporting the full lifecycle from scaffolding to production operations.
Cloud provider operations use native Azure, AWS, and GCP integration for resource provisioning, identity management, and cluster creation — no separate cloud CLI tools required.
Key Capabilities
- Zero Dependencies: Single binary — no separate kubectl, helm, or container CLI required
- Kubernetes-Native: Native Kubernetes integration for pod management, log streaming, scaling, and health checks
- Helm Integration: Full chart operations — install, upgrade, rollback, template, and chart management
- Container Operations: Build, tag, and push container images natively
- Multi-Cloud: Native Azure, AWS, and GCP integration for cloud resource management
- Agent Bundle Lifecycle: Create, build, deploy, and monitor agent bundles from a single tool
- Cluster Management: Initialize, configure, and destroy Kubernetes clusters
- Environment Promotion: Promote agent bundles through dev, staging, and production with gates
- Operations Dashboard: Real-time status, log streaming, pod restarts, and scaling from the terminal
- Dual Build Modes: Same codebase produces
ff-cli(developer-facing) andff-admin(admin operations via build tag)
curl -fsSL https://firebrandanalytics.github.io/ff-website/install.sh | sh
irm https://firebrandanalytics.github.io/ff-website/install.ps1 | iex
ff-cli --version
ff-cli skills install # Install FireFoundry skills for Claude Code / Cursor
ff-cli --help
To obtain a license, contact Firebrand Sales.
Command Groups
Commands are organized into logical groups covering the full platform lifecycle — from creating an application to deploying agent bundles to managing day-2 operations.
Each group maps to a manager in the internal layer that handles orchestration, validation,
and error reporting. Run ff-cli <group> --help to see
all available subcommands and flags.
application
Create and register applications.
ff-cli application createff-cli application register
agent-bundle
Full lifecycle — create, build, deploy, logs, status.
ff-cli agent-bundle create my-agentff-cli agent-bundle buildff-cli agent-bundle deploy
cluster
Kubernetes cluster management — init, status, destroy.
ff-cli cluster init --provider azureff-cli cluster statusff-cli cluster destroy
environment
Create, list, configure, and destroy environments.
ff-cli environment create devff-cli environment listff-cli environment destroy
ops
Operational commands — status, logs, restart, scale.
ff-cli ops statusff-cli ops logs --followff-cli ops restart
config
Configuration management — show, set, profiles.
ff-cli config showff-cli config set <key> <value>ff-cli profile list
Architecture
ff-cli integrates directly with platform SDKs — no CLI shelling
How it works: ff-cli integrates natively with Kubernetes, Helm, container registries, and all three major cloud providers. Commands go through a clean three-layer architecture: the CLI layer parses arguments, the manager layer orchestrates business logic, and the executor layer talks directly to APIs. No subprocess calls to external tools — everything runs in-process with proper error handling, retries, and structured output.
Agent Bundle Workflow
Scaffold a new agent bundle, build it into a container image, deploy to an environment, and monitor it — all from a single tool.
The agent-bundle create command generates
the project structure, Dockerfile, Helm chart, and configuration files from embedded templates.
Building and deploying use native container and Helm integration — no external tools required.
# Create a new agent bundle project
ff-cli agent-bundle create my-assistant --template hello-world
cd my-assistant
# Build the container image
ff-cli agent-bundle build
# Deploy to development
ff-cli agent-bundle deploy --env dev
# Stream logs
ff-cli ops logs --env dev --follow
# Check status
ff-cli ops status --env dev
Cluster & Environment Management
Stand up infrastructure, create isolated environments, and promote agent bundles through your deployment pipeline.
Cluster initialization provisions the Kubernetes cluster through the appropriate cloud provider SDK, configures networking, and installs the FireFoundry platform components via Helm. Environments are isolated namespaces with their own secrets, configurations, and resource quotas.
# Initialize a new cluster
ff-cli cluster init --provider azure --region westus2
# Create environments
ff-cli environment create dev
ff-cli environment create staging
ff-cli environment create prod
# Promote through environments
ff-cli agent-bundle deploy --env dev
# ... test, validate ...
ff-cli agent-bundle deploy --env staging
# ... test, validate ...
ff-cli agent-bundle deploy --env prod
ff-admin
Same codebase, different build tag. ff-admin exposes additional administrative commands for platform operators: service management, access control, audit logging, and cluster-level configuration. Built with a single flag:
# Build the admin binary
go build -tags admin -o ff-admin ./cmd/ff-cli
# Admin-only commands
ff-admin access-control list
ff-admin audit logs --since 24h
ff-admin deploy service broker --env prod
ff-admin cluster edit --scale-pool system=3
Admin commands are compiled out of the standard ff-cli binary entirely.
Go build tags ensure the admin surface area — access control, audit, deployment management, and
cluster editing — never ships to developers. One repo, two binaries, zero duplication.
Use Cases
Rapid Development
Scaffold a new agent bundle from built-in templates, run it locally with hot reload, and iterate fast with real platform services. Go from zero to a running agent in under a minute. The embedded template engine handles project structure, Dockerfiles, and Helm charts automatically.
CI/CD Integration
Scriptable commands for automated build, test, and deploy pipelines. The single binary with zero dependencies makes it trivial to add to any CI runner — just download and run. No tool installation matrix, no version conflicts, no runtime prerequisites.
Environment Management
Manage multiple environments with promotion gates and configuration isolation. Move agent bundles through dev, staging, and production with confidence that each environment is independently configured. Environment-specific secrets and broker configurations are handled natively.
Platform Operations
Monitor, troubleshoot, and manage running agents from the terminal. Stream logs in real time, check deployment status, restart pods, and scale workloads without leaving your shell. The ops commands talk directly to the Kubernetes API for immediate, accurate results.