1.4 is here · Read the release notes →

One CLI for your entire dev lifecycle

Scaffold, build, test, review, and release from one binary.
Structured JSON for AI agents. Pretty output for humans.

cargo install fledge
Get Started

Stable v1.4 release. Single binary. Cross-platform. MIT licensed.

Disconnected tools. Fragmented workflow.

Before

  • Makefile for tasks
  • shell scripts for releases
  • Separate AI tools for review
  • Tribal knowledge for onboarding

After

$ fledge run test
$ fledge review
$ fledge release patch

One config. One interface. Everything connects.

From zero to shipped

fledge templates

Scaffold from built-in or community templates. Post-hooks, git init, dependency setup. Reproducible from day one.

Templates docs →

fledge run

Auto-detects Rust, Node, Go, Python, Swift, Ruby, and Java. One command for build, test, lint. No Makefile required.

Tasks & lanes docs →

fledge spec

Keep specs and code in sync. spec check fails CI before your API breaks production.

Spec docs →

fledge review & ask

Spec-aware AI review via Ollama or Claude. Panel critiques with --with-model. Ask questions about your codebase with fledge ask.

Review & ask docs →

fledge release

Pick a bump (patch/minor/major); fledge writes the changelog from conventional commits, tags, and pushes.

Ship docs →

fledge plugins

Extend with curated defaults or community plugins. Composable lanes for custom workflows.

Plugins docs →

One binary. Three interfaces.

Fledge adapts its output to whoever (or whatever) is calling it.

Developer

Interactive 
$ fledge review
⠋ Reviewing changes against main [claude]
⚠ src/auth.rs:54: unhandled Result
✓ 7 files clean

Colored output, inline findings, spec-aware analysis.

AI Agent

Machine 
$ fledge review --json
{"schema_version": 1,
 "action": "review", ...}
$ fledge run test --json
{"schema_version": 1,
 "action": "run_task", ...}

Strict JSON, no prompts, machine-parseable. Your agents finally understand the codebase.

New Hire

Onboarding 
$ fledge doctor
✅ fledge config loaded
✅ repository — remote configured
✅ Active provider: ollama
$ fledge run dev

Guided setup, env checks, zero tribal knowledge. New hires push code on day one.

Same tool. Different flags.

Built for teams, not just individuals

Standardized Config

One fledge.toml per repo. Every project runs the same way. No more “works on my machine.”

fledge.toml reference →

Spec Enforcement

Specs drift? CI catches it. fledge spec check ensures docs and code never diverge.

Spec docs →

Fast Onboarding

fledge doctor validates the entire environment. fledge templates init my-app scaffolds with team templates.

Doctor docs →

Code Metrics

LOC, test ratios, churn. Install default plugins with fledge plugins install --defaults, then fledge metrics works immediately.

Plugins docs →

Extend without lock-in

Fledge plugins are standalone executables distributed via GitHub. Curated defaults for the common case, discover more with fledge plugins search.

# Install the curated plugin set (github, deps, metrics)
$ fledge plugins install --defaults
# Use plugin commands immediately
$ fledge metrics
# Or compose custom workflows
$ fledge lanes run ci

Get started in 60 seconds

Install

cargo install fledge

or brew install CorvidLabs/tap/fledge
or curl -fsSL https://raw.githubusercontent.com/CorvidLabs/fledge/main/install.sh | sh

Scaffold

fledge templates init my-app

or run in an existing project

Ship

fledge release patch

bump, changelog, tag, push — one shot

Read the Docs → View on GitHub