Code review built for AL & Business Central.
PEQY is the gate that lets BC teams write code at AI speed without losing trust in what ships.
Findings you can trust.
The same tools your developers use
PEQY reviews with Microsoft's AL toolbox at hand: the alc compiler and its built-in analyzers. It does not guess, it checks. With the same tools your developers use.
BC Native
PEQY resolves base app, system app and platform symbols instead of guessing from the diff. A diff alone does not contain the base app, system app or platform.
Runs where BC partners live
GitHub, GitLab and Azure DevOps, and even TFVC. Hosted in Germany; model calls go to our model provider under a no-training agreement.
See PEQY in action.
PEQY catches bugs, asks smart questions, and approves when everything looks good.
Catches critical bugs before they ship
The defects that hurt in an ERP (wrong postings, race conditions, silent data corruption) are exactly the ones that are invisible in a quick diff read. PEQY digs for them on every pull request.
Real review on a public PR: bc2adls demo PR #3. The bug shipped upstream and was later fixed in #428.
Asks smart questions on suspicious changes
Not every issue is a bug. PEQY flags changes that look intentional but risky, and asks for clarification instead of blocking.
Real review on bc2adls demo PR #2.
Approves and gets out of the way
When the code looks good, PEQY approves quickly so your team can keep shipping without bottlenecks.
Or read the full reviews yourself: PEQY on real, public pull requests at bc2adls PR #2 and PR #3.
What you'd use otherwise.
The honest comparison, by team size, including where you don't need us.
Small teams
Claude Code and your own rules
Honestly, that works while one person can hold the codebase and the rule set in their head. If that's you, you may not need PEQY yet. Come back when the team grows.
The full comparison →Growing BC partner teams
This is the gap PEQY exists for
An assistant per developer doesn't add up to a team review gate. Generic AI review wasn't built for AL. And building your own means someone on your team owns the prompts, the drift and the false positives, forever.
The honest build-vs-buy answer →Large organizations
GitHub Copilot, already bundled
Already paid for and politically easy, but not built for AL. We measured it on real Business Central pull requests: every compile error it claimed was refuted by the AL compiler.
What we measured →What BC teams say.
“Our goal was less noise and more qualified feedback. And that is exactly what happened. We configured our repositories so Copilot is excluded and only PEQY comments on pull requests.”
“It feels more like a teammate giving a review than some machine sitting there. The comments are genuinely high-quality and easy to act on.”
“Clearly better than Copilot. It feels like it really understands the language. It reviews not just the architecture, but whether we follow the AL standard.”
Answers to frequently pecked questions
Which platforms does PEQY integrate with?
PEQY reviews pull requests on GitHub, GitLab (cloud and self-hosted) and Azure DevOps, and even TFVC. Install it, open a pull request, and the review appears without changing your development process.
Why a reviewer specifically for AL and Business Central?
General-purpose AI reviewers are trained on oceans of Python and TypeScript and puddles of AL. PEQY is built for AL only: it resolves Business Central base app, system app and platform symbols, and verifies claimed compile errors against the actual AL compiler before posting them.
Couldn't we build this ourselves with Claude and some prompts?
You could. The parts are lying around: Claude, an AL MCP, the BCQuality knowledge base, some prompts. The demo takes a weekend. The product is knowing when the output is wrong: PEQY checks every claimed compile error against the AL compiler, resolves symbols against the real app layers, and gates findings on whether they can be proven, and nobody on your team has to maintain it.
Our AL is increasingly written by AI. Why does that need a separate reviewer?
Because the assistant that wrote the code can't be the independent check on it. When AI writes a growing share of your AL, review stops being optional. And reviewing your own model's output with the same model is not an independent review. PEQY is the second pair of eyes that didn't write the code.
How do you handle our code, security and privacy?
PEQY is operated by qodev GmbH in Germany and hosted in the EU; model calls go to our model provider under a no-training agreement. Analysis runs in isolated environments and your source code is not stored permanently. Your code is never used to train models, and nothing learned from your repositories is shared with other customers. Data processing details are on our legal page.
Can PEQY follow our team's coding standards?
Yes. PEQY supports custom review rules per repository, so team-specific standards and conventions are enforced alongside its built-in AL review.
Get started in three steps.
Start using PEQY now and catch serious bugs before they ship
Connect
Connect PEQY to GitHub or GitLab in a couple of clicks. Azure DevOps is set up with you by a founder.
Pull request
Open a PR and PEQY automatically analyzes your commit.
Merge
Review PEQY's suggestions and merge with confidence.