# Adaptl

> Adaptl keeps you in the loop while AI builds your product. You describe a change, Adaptl has your own coding agent (Claude Code, Codex, Aider, or any agent you mention on a GitHub pull request) build it, and you try it in a preview full of realistic, made-up customers. You pin what's wrong, the agent fixes it, and your team merges when it works.

Key facts:

- For: solo builders and small teams who ship with AI coding agents, and product teams (product managers, tech leads, founders, agencies) who want to try a change instead of reading the diff.
- Works with your GitHub repository, or a project started from scratch. Adaptl opens the issue, branch and draft pull request for each change.
- You keep your own coding agent and AI key. Adaptl writes the brief and checks the result; the agent writes the code.
- Previews are filled with made-up companies, people and history, written through the app's own models (Django, Prisma, SQLAlchemy, or SQL) on PostgreSQL. MySQL is planned.
- Never uses production data. Adaptl reads code and database structure, not customers' records.
- Feedback is pinned on the running app, tied to the screen and the customer you were signed in as, and goes back to the agent as one batch.
- A plain-English history of how the product evolved, grouped into chapters, built from merged pull requests.
- A living spec without writing one: before a change merges, Adaptl lists what it changes for users (new, works differently, goes away), each line based on the change itself and reviewed by the team; merged changes build an up-to-date description of what the product does.
- Review links that open one change for a client; they expire and can be withdrawn.
- Helper agents work beside the coding agent: speed and memory (working today), and errors, infrastructure cost and code review (coming). Each starts from a shared expert version, remembers what it learns about your project, and can be adjusted for your team.
- People always merge. Adaptl prepares a normal GitHub pull request.
- Runs online at https://inloop.adaptl.ai (hosted in the EU) or on your own computer. Signing up is free; you use your own AI key.
- Made by Adaptl, registered in the Netherlands (no. 42162373).

## Pages

- [Home](https://adaptl.ai/): what Adaptl does, how it works, who it's for, questions and answers
- [Full text for AI assistants](https://adaptl.ai/llms-full.txt): the home page as plain Markdown
- [Compare](https://adaptl.ai/compare.html): Adaptl next to app generators, coding agents, preview deployments and named tools, with sources
- [Privacy](https://adaptl.ai/privacy.html): what data Adaptl handles and why
- [Sign in or start free](https://inloop.adaptl.ai): the app

## Blog

- [How to review AI-written changes without reading the diff](https://adaptl.ai/blog/review-ai-changes-without-reading-code.html)
- [Empty preview environments lie](https://adaptl.ai/blog/synthetic-personas-preview-environments.html): why realistic, related test data catches what screenshots miss
- [Product lineage for AI codebases](https://adaptl.ai/blog/product-lineage-for-ai-codebases.html)
- [The staging tax](https://adaptl.ai/blog/staging-tax.html): what broken shared staging costs a team that ships with AI