Stop re-explaining your company to Claude.

One git repo. Every AI session starts with your product, stack, customers, and team knowledge. Automatically.

$ npx @antidrift/cli init
No sign-up. No vendor. One command and every AI tool knows your company.

Use Cowork or Claude Desktop? Antidrift works there too. Get started →

The problem

Your AI tools don't know your company.

Every session starts from zero

You open Claude Code and spend the first 10 minutes re-explaining your product, your stack, your customers, how you talk about things. Every session. Every teammate. Every machine.

Context files get passed around

Someone writes a context file with company knowledge. It gets copy-pasted into Slack. Three versions exist. None of them are current. The new hire doesn't have any of them.

Context drifts across the team

Your positioning changed last month but half the team's prompts still use the old version. Everyone's Claude is working from slightly different context.

New hires start from scratch

You spent months building context in your AI sessions. A new teammate joins and their Claude knows nothing. Onboarding means re-explaining everything from zero.

antidrift is the fix.

One git repo that gives every AI tool full context on your company. Clone it, open your tool, and it already knows everything. The whole team shares the same brain. It stays in sync through git.


The fix

Your company knowledge loads automatically.

Every CLAUDE.md, AGENTS.md, and GEMINI.md in your repo is read by every AI session. No prompting. No pasting. Your whole team gets the same context on every machine.

Your brain
product/
CLAUDE.md · AGENTS.md · GEMINI.md
engineering/
CLAUDE.md · AGENTS.md · GEMINI.md
marketing/
CLAUDE.md · AGENTS.md · GEMINI.md
sales/
CLAUDE.md · AGENTS.md · GEMINI.md
Every session starts here
Any teammate. Any AI tool. Full context. Day one.
Learn more: Auto-Context →

How it works

Context. Skills. Connections.

Three layers that make your AI agent actually useful.

Skills

Slash commands your team shares. /prep, /review, /recap - build your own or install packs.

Live Connections

19 connections, 150+ tools. Stripe, Google, GitHub, Attio - your agent reads and writes directly.

Cross-Platform

One brain for Claude Code, Codex, and Gemini. Write once, runs everywhere.

Plain Git

No database. No vendor. Just git. Branch it, diff it, own it.

Explore all skills →

Connections beta

One command. Live data.

Connect your services. Your AI agent reads and writes directly - no glue code. Works with Claude Code, Codex, and Gemini.

Popular

Google Workspace

Sheets, Docs, Drive, Gmail, Calendar - 37 tools

Attio CRM

People, companies, deals, tasks, notes - 14 tools

GitHub

Repos, issues, PRs, actions, releases - 15+ tools

Stripe

Customers, invoices, subscriptions, charges - 17 tools

New

New Linear

Projects, issues, cycles, teams - 12 tools

New Vercel

Projects, deployments, domains, env vars - 9 tools

New Cloudflare

DNS, Pages, Workers, R2 - 12 tools

New Pipedrive

Deals, contacts, organizations - 15 tools

Finance
Browse all connections →

Works with every tool

Your team shouldn't have to pick one AI tool.

Engineers on Claude Code. Teammate trying Codex. Contractor on Gemini. One brain works in all of them - context and skills translate automatically.

No tool mandates

Everyone uses what they prefer. The brain translates context and skills to each tool's native format. No one switches tools, no one misses context.

Always the same brain

Update the brain and every platform sees the change. It's maintained by working inside it - not by someone keeping docs current.

No lock-in

Plain markdown in a git repo. No vendor, no database, no account. Switch tools tomorrow - the brain comes with you.


Your next Claude session starts with full company context.

No sign-up. No API key. Works with Claude Code, Codex, and Gemini.

$ npx @antidrift/cli init