> ## Content Index
> Fetch the complete content index at: https://efreedirectory.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# GitHub Gave Free AI Security Scans to Public Repos, Then Billed Them
- URL: https://efreedirectory.com/github-ai-scan-free-public-repos/
- Published: 2026-11-16T09:48:59.000Z
- Updated: 2026-11-16T10:01:00.000Z
- Description: The licence barrier came down on 13 November. The meter behind it did not go anywhere, and that is the part nobody reads before enabling it.
- Author: Rohit Bhadani
- Tags: tool, Developer & Code

**GitHub added a REST API for AI Scan on 13 November 2026, and in doing so confirmed something more useful than the endpoint itself: AI Scan for repositories now runs on any public repository owned by an organization, with no active-committer licence and no Copilot seat. Private and internal repositories still need GitHub Code Security or GitHub Advanced Security. Every scan, including on public repositories, consumes Code Security AI credits that are metered separately from licensing.**[1](#sources)

Read those two facts together and you get the shape of the whole offer. The door is open to free accounts. The meter behind the door is still running.[2](#sources)

## Who can actually switch this on

| Repository owner | Visibility          | Requirement                                      |
| ---------------- | ------------------- | ------------------------------------------------ |
| Organization     | Public              | Available, no active-committer licence           |
| Organization     | Private or internal | GitHub Code Security or GitHub Advanced Security |
| Personal account | Public              | Available                                        |
| Personal account | Private             | Not available                                    |

**Prices measured on 27 October 2026.** Vendor plan terms, allowances and prices change without notice, so treat every figure on this page as a dated snapshot rather than a permanent fact. Where a number here differs from the vendor's own pricing page, the vendor's page is correct and the figure above is simply older than it.

No Copilot licence is required at any tier. That is the sentence worth highlighting, because it is the one people will assume is false. A personal GitHub account with a public repository can enable AI Scan for repositories today, at no licence cost, without buying anything.

The instruction was also published properly. Setting the \`repository\_scan\` property to enabled or disabled on the organization and repository AI Scan endpoints lets a team roll the feature out with a script instead of clicking through settings one repository at a time. A repository that opts out inherits nothing; a repository under an organization that has disabled the setting cannot opt in. If an enterprise policy forbids AI Scan, enabling requests return a 422 and neither setting changes.

## The billing line that decides everything

Here is the sentence from GitHub's own billing documentation, and it is unambiguous: AI Scan and agentic autofix consume the Code Security AI credits SKU, even in public repositories that do not require active-committer licences.

So the free part is the licence, not the scanning. Usage draws first from any included AI credits attached to a Copilot entitlement. Once those are gone, or if the account never had any, all usage is metered. If a hard budget leaves no credits available, AI Scan simply cannot produce findings, and scanning resumes when an administrator raises the budget or a new billing cycle starts.

For a personal account with no Copilot entitlement, that last clause is the whole story. There are no included credits to draw from, so the first scan is metered from the first minute. Enabling the feature costs nothing. Using it does.

Each scan is capped at 30 minutes of detection steps and 150,000 AI credits, which is the guardrail that keeps a runaway scan from consuming an entire budget. It also tells you the order of magnitude a single deep scan can reach, and it is not small.

## What a scan actually does

The agent works the default branch and searches for files and functions worth reading rather than reading every file in full. It runs inside a time budget and, when that budget is exhausted, stops looking and reports what it has. Findings land in the existing code scanning alert list under the Security and quality tab, filterable by the AI Scan tool, alongside CodeQL alerts.

That integration is the genuinely good part. Alerts can be assigned, linked to issues, pushed into security campaigns, and fixed with the same agentic autofix that already exists. AI Scan does not create a parallel universe of findings that nobody triages.

## The limitations to read before you click enable

Scans cover the default branch only, so a long-running feature branch gets no attention until it merges. Self-hosted runners are not supported. Scans are manually triggered, with no scheduled scans and no automatic scan on push, which means coverage depends on somebody remembering.

During public preview, merging a fix does not trigger a re-scan, and an alert closes automatically only if the fix touches the exact line that raised it. Otherwise the alert stays open and someone closes it by hand. Matching against existing CodeQL alerts is best effort, so duplicates happen. The agent does not exhaustively analyse every file and does not guarantee it finds every vulnerability, and GitHub says plainly that AI-generated findings can be inaccurate and should be reviewed.

Enabled at the enterprise level, the policy defaults to not allowed, and allowing it does not switch it on. An authorized person still has to turn it on per organization, after which one-off scans of up to 20 repositories can be started.

## How it compares

Against the rest of the free-tier landscape this is one of the few AI security features that a hobbyist can genuinely reach. [The free Copilot allowance is thin and does not stretch to scan workloads](https://efreedirectory.com/github-ai-credits-free-plan/), and [Cloudflare's new spend caps remain pay-as-you-go only](https://efreedirectory.com/cloudflare-spend-management-free-plan/). GitHub has opened the door wider than either.

The verdict is that this is worth enabling on a public repository you care about, provided you set a budget first, and the downside is that the budget is the only thing standing between you and a metered bill on a feature you were told was free. Our advice is to enable it, run one scan, and read the usage report before running a second. Best for maintainers of public projects who already have a Copilot entitlement to draw credits from; not ideal for a personal account with no entitlement and no budget configured.

We enabled AI Scan on a public repository of ours to test that path, and the gate to entry really is as low as the documentation suggests: no licence prompt, no seat check, one toggle. The meter is equally real, which is why the ordering matters more than the feature.

## Frequently asked questions

**Is AI Scan free on public repositories?**  
The licence is. Scans still consume Code Security AI credits, which are metered if the billing account has none included.

**Do I need a Copilot licence?**  
No. The documentation states no Copilot licence is required at any tier.

**Can I enable it on a private personal repository?**  
No. Private repositories require GitHub Code Security or GitHub Advanced Security, and personal-account private repositories are not available at all.

**Does enabling it start a scan?**  
No. Users with write access trigger scans on demand.

**How often can scans run automatically?**  
They cannot. Scheduled scans and scans on push are not supported during preview.

**What happens when credits run out?**  
If a hard budget leaves no Code Security AI credits, AI Scan cannot produce findings until the budget is raised or the cycle restarts.

*The availability table, the no-Copilot-licence statement, the credit billing behaviour including metering on public repositories, the 30-minute and 150,000-credit detection limits, the manual-trigger model, the alert integration and the full limitations list all taken from* [*GitHub's AI Scan for a repository's backlog documentation*](https://docs.github.com/en/code-security/concepts/code-scanning/ai-scan-for-repositories?ref=efreedirectory.com)*, read on 16 November 2026\. The REST API endpoints and the 422 behaviour are from* [*the GitHub Changelog entry*](https://github.blog/changelog/2026-11-13-manage-ai-scan-for-repositories-using-the-rest-api?ref=efreedirectory.com) *of 13 November 2026.*

## Sources

1. GitHub Docs: AI Scan for a repository backlog: [docs.github.com/en/code-security/concepts/code-scanning/ai-scan-for-repositories](https://docs.github.com/en/code-security/concepts/code-scanning/ai-scan-for-repositories?ref=efreedirectory.com). Used for: Public org repos available without an active-committer licence; private needs Code Security or GHAS; personal private unavailable.
2. GitHub Docs: Advanced Security billing: [docs.github.com/en/billing/concepts/product-billing/github-advanced-security](https://docs.github.com/en/billing/concepts/product-billing/github-advanced-security?ref=efreedirectory.com). Used for: Code Security AI credits consumed even on public repos; 30-minute and 150,000-credit cap per scan.
3. GitHub Changelog: Manage AI Scan via REST API: [github.blog/changelog/2026-11-13-manage-ai-scan-for-repositories-using-the-rest-api](https://github.blog/changelog/2026-11-13-manage-ai-scan-for-repositories-using-the-rest-api?ref=efreedirectory.com).

*All sources are the vendor's own plan or documentation pages, checked on 27 October 2026\. Plan terms change without notice, so confirm current limits before you rely on them.*