Up-to-date developer docs for AI agents to write better code

playbooks gives better context to Claude Code.

Stop hallucinations and bad code by giving AI agents up-to-date context from public docs or your own private repos.

How it works

Add the playbooks MCP to your AI coding tool. When your agent needs documentation, it searches, fetches what it needs, and keeps coding.

01

Your agent searches

Semantic search across public libraries or private GitHub repos. Returns code snippets with line ranges.

02

Fetches code snippets

Agent fetches specific snippets using the line range. No full docs dumped - just the relevant code and context.

03

Fetches full docs if needed

If your agent needs more, they can remove the line range and fetch the full document context only if necessary.

Public docs
GitHub, doc sites
Private docs
GitHub, doc sites, PDFs
playbooks crawler
playbooks MCP
search
"how do I..."
fetch
get code snippets
IDEs, Chats, Agents

Less context, better code

playbooks uses 90%+ less tokens by giving your agent only the examples it needs, without overloading your context.

04

Progressive disclosure

10k+ tokens dumped into your context costs you money and makes your agent less effective.

Using progressive disclosure, your agent gets matching code snippets and can fetch more context only when it needs it.

05

Avoid usage limits

Every token your agent processes costs money in real API spend or by reducing your agent usage limits.

Less tokens wasted means you can code for longer with the same accurate, up-to-date context you need.

06

Docs always up-to-date

A lot of tokens get wasted when you get stuck in the "circle of death" and your agent can't figure it out.

playbooks finds and updates documentation automatically so your agent never has to guess how to code that new feature.

Add playbooks to your agent

Create an account, grab your API key, and paste the config below. Your agent gets three new tools.

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
    "mcpServers": {
        "playbooks": {
            "url": "https://playbooks.com/api/mcp",
            "headers": {
                "Authorization": "Bearer YOUR_API_KEY"
            }
        }
    }
}
07

search_documentation

Search public libraries or your private repos. Returns matching snippets with line numbers.

08

fetch_documentation

Fetch specific lines from any search result. Get exactly what you need, nothing more.

09

fetch_url_markdown

Fetch any URL as clean markdown. Great for one-off urls you need to give the agent as context.

Simple pricing

Pick a plan that fits. Searches cost 1 credit each. Fetch tools are free. Credits refresh monthly.

Mini
Start here
500 credits / month1 seat · 1 project
$5 / month
Plus
2,000 credits / month3 seats · 3 projects
$15 / month
Pro
5,000 credits / month5+ seats · 10 projects
$25 / month

Common questions

Everything you need to know about getting started with playbooks.

01

What can I search?

Thousands of public libraries and frameworks - Laravel, React, Next.js, Tailwind, and more. Docs are indexed and updated automatically.

You can also connect your private GitHub repos via OAuth and search those alongside public docs.

02

How do credits work?

Searches cost 1 credit each. Fetch tools are free - your agent can grab what it needs without extra cost.

Credits refresh monthly. Unused credits don't roll over.

03

How is this different from Context7?

Context7 returns large blocks of documentation - often 10,000+ tokens per call. That's a lot of context for your agent to parse.

We return focused snippets with progressive fetching. Smaller context means sharper agents.

04

What tools does this work with?

Any MCP-compatible client. Claude Code, Codex CLI, Cursor, Factory, Amp, Zed, and more.

If your tool supports the Model Context Protocol, it works with playbooks.

05

Can I use private repos?

Yes. Connect your GitHub account via OAuth in the dashboard, then add repos to your project.

We index your code automatically so your agent can search it alongside public docs.

06

Why don't you have a free plan?

I'm a solo founder, not venture-funded. Servers, crawlers, and indexing cost real money to run and maintain.

At $5/month, the Mini plan is about as close to free as it gets.

Give your agent better context.

Get better code - no hallucinations.