AI content moderation platform with review queues, appeals and EU DSA reporting built in, moderating text, images, video and audio in 100+ languages.
Pick Checkstep if you run a European marketplace, dating app or game community with a small safety team and need DSA and Online Safety Act compliance done for you. Skip it if you only want a classifier to call: there is no self-serve plan. Checkstep says most teams on it have five people or fewer.
It files the legally required explanation of each removal to the EU's DSA database by itself.
Once a platform's DSA statements of reasons, appeals and transparency reports run through Checkstep, moving them is a compliance project, while its reliance on partner models leaves detection quality partly in other hands.
Main trade-off
No way to try or price it without a sales conversation.
Checkstep does not publish prices. Access starts with a booked demo, and contracts are agreed with sales.
Checkstep is a content moderation platform from a UK company founded in 2020. It scores text, images, video and audio, routes borderline items to moderators and handles the EU Digital Services Act paperwork, filing statements of reasons and running appeals. It acquired Unitary's trust and safety business in November 2026.
Multimodal scanning
Moderates text, images, video and audio and reads them together, so an image is judged with its caption, in more than 100 languages natively.
Per-policy confidence thresholds
Each policy has its own threshold, so clear violations are removed automatically and borderline items are routed to a moderator queue.
DSA statements of reasons
Every enforcement records the rule broken and is filed to the EU DSA Transparency Database automatically, with transparency reports produced from the same log.
Appeals workflow
Users can appeal a decision and a human reviews it, which the DSA requires and the platform tracks end to end.
Partner models
Voice analysis from Modulate and OpenAI's CSAM image classifier plug into the same queues as Checkstep's own detection.
Delegated media access
Images and video are fetched by link rather than copied into Checkstep, which keeps customer media out of its storage.
| Tool | Price from | Free plan | Runs on | Rating |
|---|---|---|---|---|
| Quote | No | Web | 3.8 | |
| Quote | No | Web | 3.6 | |
| Quote | No | Web | 3.9 | |
| Quote | No | Web | -- |
Proprietary
Checkstep keeps its prices private. You book a demo, describe your volumes and the content types you need covered, and sales returns a contract quote.
No free plan or self-serve trial is listed on its website. The route in is a booked demo.
Checkstep announced in November 2026 that it had acquired Unitary's trust and safety business. Its announcement says Unitary customers keep their contracts and moderators and move onto one platform for AI and human moderation.
Yes, that is its main selling point. It records which rule each removal enforced, runs a human-reviewed appeal process, submits the required explanation of every decision to the EU database and builds the transparency report. Policies can differ by country.
Checkstep says it does not store your content: media is read by link through delegated access. Its security statement adds that customer data is encrypted at rest and used only to serve API calls.
Checkstep is a London-registered moderation platform founded in 2020 and run as a fully remote company, led by chief executive Guillaume Bouchard. It sells one place to scan user content, enforce a written policy and prove compliance, and in November 2026 it announced that it had acquired the trust and safety business of Unitary, a fellow UK moderation vendor, with Unitary's customers keeping their contracts and moderators.
The platform mixes several detection models, its own and partners', and sets a confidence threshold per policy: confident detections are actioned automatically and uncertain ones go to a human queue. It reads an image and its caption together, so a harmless photo with a hateful caption is still caught. Media is fetched by link through delegated access rather than stored, which Checkstep says shortens a customer's data-protection review. Voice comes from a partnership with Modulate, whose voice analysis feeds Checkstep's queues, and Checkstep says it has access to OpenAI's CSAM image classifier for child-safety detection.
What sets it apart is the compliance layer. The EU Digital Services Act and the UK Online Safety Act require a platform to tell users which rule they broke, offer a human-reviewed appeal and publish transparency reports. Checkstep records the violated policy on every action, runs the appeal workflow, files statements of reasons to the EU DSA Transparency Database automatically and produces the report. That suits marketplaces, dating apps, gaming communities, streaming services and review sites in Europe with small trust and safety teams; Checkstep says most in-house teams using it are five people or fewer.
It is more platform than many buyers need. A developer who only needs a classifier can call Deploy Content Safety per request, or a free endpoint from OpenAI, and wire up their own rules. A very large platform facing organised abuse may want the threat intelligence of ActiveFence, and a team that already has detection and only needs case management and policy tooling can compare Cinder. Hokai files all of these under AI governance tools.
Product facts on this page (pricing, plans, integrations, compliance) are read from Checkstep's own pages and documentation. The verdict is Hokai's reading of them, not the vendor's claim.