> For the complete documentation index, see [llms.txt](https://tools.osintnewsletter.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://tools.osintnewsletter.com/osint-tools/deepware.md).
# Deepware
| **Deepware** | **Quick Overview** |
| ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| URL | |
| What it does | Detects potential deepfakes in video and audio using machine learning models trained to identify signs of synthetic media manipulation. |
| How to use it | Upload a video or provide a supported video URL, then review the confidence score indicating the likelihood that the content contains AI-generated or manipulated elements. |
| Cost | Partially Free (free web analysis with additional enterprise capabilities available). |
| Account required | No (for basic web analysis). |
| Cookies | Standard analytics cookies (Google Analytics) and an essential Cloudflare security cookie. |
| Ownership | Founded in Sarajevo in 2018 by [Akif Iris](https://platform.tracxn.com/a/d/people/58b3aafde4b072d103006493/akifiris) and [Orhan Akyurek](https://platform.tracxn.com/a/d/people/58b3aa4ce4b072d103006356/orhanakyurek) and Launched as a research project by Zemana. |
| Use in Reporting | Useful for assessing the authenticity of suspected deepfake videos before publication, supporting verification workflows during breaking news, disinformation investigations and digital forensic analysis. |
### What does Deepware do?
Deepware applies multiple detection models to uploaded media, looking for artefacts commonly associated with deepfake generation techniques such as facial synthesis, face swapping, lip synchronisation and AI-generated speech.
Results are presented as a confidence assessment rather than a definitive conclusion. (As deepfake generation techniques evolve rapidly, no automated detection tool can guarantee complete accuracy.)
**The lowdown:** It’s best used as an initial screening tool within a wider verification process.
### How to Use:
**1. Upload a video file or submit a supported video URL for analysis.**
**2. Review the detection results and confidence score, paying attention to which detection models flagged potential manipulation.**
**3. Corroborate findings using additional forensic techniques such as metadata analysis, reverse image searches, source verification and visual inspection before drawing conclusions.**
### Cost
* [ ] Free
* [x] Partially Free
* [ ] Paid
Free web analysis with additional enterprise capabilities available.
## Data Processing
### Account Required:
* [x] Yes
* [x] No
No for basic web analysis. Yes for full functionality.
### Cookies:
The site uses Google Analytics cookies (\_ga, \_gid) to collect anonymous website usage statistics and a Cloudflare security cookie (cf\_clearance) to verify legitimate users and protect against automated abuse.
### Use in Reporting
Deepware can be used to:
* Screen suspected deepfake videos before publication.
* Support verification during breaking news and crisis events.
* Assess user-generated content circulating on social media.
* Corroborate findings during misinformation and disinformation investigations.
* Assist journalists and investigators when authenticating visual evidence.
* Provide an initial forensic assessment prior to expert analysis.
Deepfake detection tools such as deepware have become part of verification workflows used by journalists, fact-checkers and OSINT investigators when analysing manipulated media relating to elections, armed conflicts, financial fraud and misinformation campaigns.
| **Capabilities** | **Limitations** |
| --------------------------------------------------------------------- | -------------------------------------------------------------------------------------- |
| Detects potential AI-generated or manipulated video content. | Detection results are probabilistic and not definitive proof of manipulation. |
| Analyses uploaded videos and supported URLs. | Accuracy varies depending on the deepfake generation method and media quality. |
| Applies multiple AI detection models to improve detection coverage. | May produce false positives or false negatives. |
| Produces confidence scores indicating the likelihood of manipulation. | Limited effectiveness against newly developed or highly sophisticated deepfake models. |
| Provides rapid screening as part of a digital verification workflow. |
|
### Summary
Deepware is most valuable after a video or audio file has been identified and before it’s relied upon as evidence/included in reporting. It provides an initial assessment of whether media may contain AI-generated or manipulated content, helping investigators prioritise material for further forensic examination.
### Ownership
According to Tracxn, Deepware is an unfunded company based in Sarajevo (Bosnia And Herzegovina), founded in 2018 by [Akif Iris](https://platform.tracxn.com/a/d/people/58b3aafde4b072d103006493/akifiris) and [Orhan Akyurek](https://platform.tracxn.com/a/d/people/58b3aa4ce4b072d103006356/orhanakyurek). Its parent company is Zemana who researched methods to develop an AI-based antivirus engine.
### Ethical Considerations
* Detection scores should not be presented as definitive evidence that media is fake.
* Avoid making accusations based solely on automated AI detection results.
* Corroborate findings through metadata, provenance, contextual analysis and independent verification.
* Consider the possibility of false positives and false negatives, particularly with compressed or low-quality media.
* Be transparent about the limitations of AI detection when reporting findings.
### Related Tools:
* Reality Defender
* Hive Moderation AI Detector
* TrueMedia.org
* Intel471 Deepfake Detection Resources
* [Forensically](/osint-tools/forensically.md)
* FotoForensics
#### Sources