Detect synthetic content before it deceives

Heimdallr is an open-source tool for detecting AI-generated text, voice, and images. It runs on Gjallarhorn One, a custom detection model built from the ground up.

Try Heimdallr
About

What is Heimdallr

Language models write text, clone voices from recordings, and generate images from prompts. This synthetic content is getting harder to tell apart from the real thing. Heimdallr is being built to help anyone make that call.

At the core of Heimdallr is Gjallarhorn One, a detection model developed specifically for this purpose. It examines content across multiple modalities and returns a clear, interpretable result, with no dependency on third-party AI services.

Heimdallr is free software, released under the GNU Affero General Public License 3.0. The source code is public, the development process is open, and the tool will remain free for individuals.

How it works

How it works

01

Submit content

Upload a file or paste text directly. Heimdallr accepts documents, audio recordings, and images.

02

Analysis

Gjallarhorn One examines the content for patterns characteristic of AI generation. No content is stored after the session ends.

03

Clear result

You get a straightforward verdict on whether the content appears synthetic, along with the key signals behind it.

Product

Why Heimdallr

Open source

Every line of code is public. Anyone can read, audit, and contribute to the project on GitHub.

Privacy-first

Content you submit is not stored, logged, or used for training. Your data never leaves the session.

Purpose-built model

Gjallarhorn One was built specifically for synthetic content detection. It is not an adaptation of a general-purpose model.

Free for individuals

Heimdallr is and will remain free to use for individuals. No subscription required, no paywall.