Know what a change can reach before it ships.

Your team ships more code than anyone can hold in their head, much of it written with AI. Ask Jentic reads your repository and shows what any change can reach: the files, tests, routes and data behind it.

  • ✓ A map built from your code, labelled extracted or AI-written
  • ✓ What a file or a whole release can reach, and which changes no test covers
  • ✓ The same answers for your team and their AI assistants

Your first run on a repo is free. No card.

Facts from your code

Ask Jentic extracts dependencies, the data model, routes and services straight from your code and history. Documents are labeled as extracted fact or AI-written (within limits).

AI IDEs usually grep for keywords and build a picture from what matches, and they can miss things and give confidently wrong advice.

See what a change can reach

For any file: what depends on it, the tests, routes and services it affects, and files that change together.

Python, Go, TypeScript and JavaScript imports are resolved to the files they load, using each language's own resolution rules. Other languages, including Rust and Java, are matched by name, so less exact.

Check a release before it ships

Compare two branches, tags or commits and see every changed file, what each one can reach, and which changes no test covers.

Built from your repository's map and GitHub's comparison, with no AI model. It shows what a release can reach. It does not predict whether it will fail.

Docs, decisions and how it runs

An overview, setup notes and decision records from your change history, plus how the project is built and run.

Runtimes and end-of-life dates, containers, CI, commands and pinned dependencies.

Ask in chat or from your IDE

Ask Jentic gives your AI the big picture in two ways: static files in your repo that any assistant can read, and a connection (MCP) that lets your AI IDE ask about your code directly. You stop re-explaining the project every time you start a session.

The files are plain Markdown, JSON and Mermaid, proposed as a pull request and refreshed on every push or on a schedule. You can also chat with your repo.

Get started

What happens to your code

  • Read-only access to the repository or repositories you choose, copied to a temporary workspace for each run and deleted when it ends.
  • Excerpts go to OpenAI's API only to write and quality-check your documentation, we don't train models on your code, and our logs keep sizes and timings.
  • We keep the generated documentation and its chat index, never your source files; changing the repository or deleting your account clears the index.
  • Docs arrive as a pull request your team reviews.
  • Incident context is optional and off until you connect your logs. Its error data never goes to an AI model.

Read the full data-handling details

Frequently asked questions

Straight answers about what Ask Jentic is, what it does, and how to start.

What is Ask Jentic?

Ask Jentic helps your team understand a codebase before they change it. It reads your GitHub project and its change history, maps how the pieces fit, shows what a change to any file or a whole release can reach, and writes documentation and short decision records. It runs on every push or on a schedule and saves the results back to your project.

Can I check a whole release before it ships?

Yes. The release check compares two branches, tags or commits and lists the changed files, what each one can reach, and which changed files no test covers. It works from your repository's map and GitHub's comparison, with no AI model and no extra GitHub permission. It shows what a release can reach; it does not predict whether it will fail.

Does Ask Jentic help when something breaks in production?

Yes, in early access. When a production error fires, responders get a short message: the files involved, what changed there in the last 14 days, and what else those files reach. A weekly brief shows where incidents concentrate, and an incident review starter gathers the facts and leaves the conclusions to your team. It works with Google Cloud logs today: the repository owner connects it from the dashboard with one setup script. It sends no incident data to an AI model.

How is Ask Jentic different from writing docs by hand?

Hand-written docs drift, and nobody writes the dependency map or the list of what a change can break. Ask Jentic bases every document on your actual code, refreshes it automatically, and leaves alone topics your own docs already cover well.

How do I get started with Ask Jentic?

Subscribe on the product page and connect your project on GitHub. Each subscription covers one project and includes a monthly allowance of runs; add a subscription for each additional project.

Is the documentation written by AI?

Partly, and every document says which part. The dependency map, data model, routes, services, operations details and change-impact tables are extracted from your code with no AI model involved. The overview and explanations are AI-written from those facts.

What does the change-impact analysis show?

For each source file: the files that depend on it, directly and indirectly, the tests that reach it, the routes and services it can affect, the data it defines, and files that often change in the same commit. It is saved as a readable page and as a data file (impact.json).

How accurate is the change-impact analysis, and which languages does it cover?

We check it against each language's own tooling: the Go toolchain and the TypeScript compiler for Go, TypeScript and JavaScript, and, for Python, by deliberately breaking files and seeing what really fails. Python, Go, TypeScript and JavaScript imports are resolved to the files they load. Rust and Java are matched by name, so some links can be missed or wrong. It does not see imports built at runtime, calls between separately deployed services, or monorepo packages imported by name.

How is this different from asking an AI assistant in my editor?

An assistant answers the question you ask, from what it reads at that moment. Ask Jentic runs on every push and saves the results as files in your project, so your team and their tools read the same answer, you can see what changed since the last run, and it stays current without anyone asking.

Is there a limit to how much one run covers?

Every run analyses your whole repository. The AI-written parts are capped: up to 30 top-level folders get a module page, largest first, and up to 20 decision records are written, most significant first. A run tells you when it hits a cap. For a larger repository, email hello@ask-jentic.ai.

How much does Ask Jentic cost?

One plan: a flat monthly price per project that includes a monthly allowance of runs, plus a fixed price for each additional run. Exact prices are on the pricing page.

Why is it called Ask Jentic?

It is named after its founder, Jen Anderson, so "Jen-tic" is a person you can ask. Don't guess how your code fits together: ask Jen-tic, or let your AI assistant ask.

Who built Ask Jentic?

Ask Jentic was founded by Dr. Jen Anderson, a technology executive specializing in AI integration and digital transformation. It is operated by Astrix8 Partners.

Don’t guess. Ask Jen-tic.

Get started