Skip to main content

AI setup tailored for your codebase.

Scores your AI setup. Generates what's missing. Syncs to your team via git.

$npx @rely-ai/caliber bootstrap
Star us on GitHub
terminal
$

The problem

Bad setup = bad agent.

Without setup

No config, stale context

  • No project context for agents
  • No learning from past AI sessions
  • Missing MCPs that unlock key features
  • Stale configs nobody updates

With setup

Full context, always up to date

  • Full project context generated
  • Session learnings captured automatically
  • Right MCPs recommended and installed
  • Configs stay fresh as code changes

The challenge

Why is it hard to have a perfect setup?

Invisible gaps in your setup

Better MCPs for your DB, skills that halve deploy time — you don't know they exist.

Best practices change weekly

New tools drop, guides update. Last month's best practice is today's tech debt.

Configs rot while code evolves

You refactor daily but CLAUDE.md still references last quarter's architecture.

Every developer has their own setup

One dev has MCPs, another doesn't. No single source of truth — until now.

repeat forever

Meet Caliber

Best practices, generated for your codebase.

Curated skills, configs, and MCP recommendations from research and the community.

Caliber Score

Deterministic. No LLM needed.

$ caliber score
Score: 0/100 Grade: D
Files & Setup 8/25 · Quality 10/25 · Grounding 5/20
Accuracy 7/15 · Freshness & Safety 5/10
$ caliber init && caliber score
Score: 0/100 Grade: A (+59 pts)
Files & Setup 24/25 · Quality 23/25 · Grounding 19/20
Accuracy 15/15 · Freshness & Safety 10/10
Quality scored against SkillsBench — the open benchmark for AI coding skills
3594+59 pts
Files & Setup
824/ 25

Configs, skills, MCPs present

Quality
1023/ 25

Benchmarked via SkillsBench

Grounding
519/ 20

References match real paths

Accuracy
715/ 15

Commands and paths exist

Freshness & Safety
510/ 10

Recency, secrets, permissions

No hardcoded API keys detected
No leaked secrets in MCP configs
Permissions explicitly set
.gitignore covers sensitive paths
Configs updated within 7 days

Commands

Simple CLI. Powerful workflow.

caliber bootstrapInstall agent skills — the fastest way to start
caliber initFull setup wizard with analysis and generation
caliber scoreRate your setup 0–100 (--compare across branches)
caliber regenerateRe-analyze and regenerate all configs
caliber refreshUpdate docs based on recent code changes
caliber skillsDiscover and install community skills
caliber learnCapture patterns from AI coding sessions
caliber hooksManage auto-refresh automation

Works with

Claude CodeNo key needed
CursorNo key needed
OpenAI CodexAPI key
GitHub CopilotNo key needed
OpenCodeAPI key

No API key needed — use your existing subscription.

Team sync

One dev sets up. Everyone benefits.

Configs live in your repo. Clone, code — same setup, every time.

First developersets up the project
via git
Another teammateclones and starts coding

Git-native distribution

Plain files in git. No sync server, no accounts, no lock-in.

Automatic freshness

Run caliber refresh, commit. Every pull brings the latest configs.

Network effect

First dev sets up once. Every future contributor gets the full config from day one.

Get started

Try it in 30 seconds.

$npx @rely-ai/caliber bootstrap

No install needed — runs instantly via npx. 100% local

FAQ

Caliber is a command-line tool that generates AI coding configs — CLAUDE.md, .cursorrules, skills, AGENTS.md, and MCP server setups — tailored to your codebase. It supports Claude Code, Cursor, GitHub Copilot, Windsurf, and works with JavaScript, TypeScript, Python, Go, Rust, Java, and more.