by Alice (formerly ActiveFence)
Alice's trust and safety service for user-generated content, pairing AI detection across 20+ abuse areas with threat intelligence and analyst teams.
Pick ActiveFence if you run a very large platform facing organised abuse, such as extremist networks or child exploitation rings, and want intelligence and investigators as well as detection. Skip it for a smaller product: it has no self-serve access or public prices. Alice has sold it under this name since 2018.
Threat intelligence feeds and human investigators come bundled with the detection; the product page says the feeds surface harmful narratives and coordinated manipulation in real time (read 22 Nov 2026).
Alice says its Rabbit Hole collection holds billions of real-world harmful samples in 120+ languages (alice.io, read 22 Nov 2026), data a newer rival cannot rebuild quickly, while the company's turn to AI security leaves room for moderation-only vendors.
Main trade-off
Nothing can be tried or priced without a sales cycle.
No public pricing. ActiveFence is sold by quote after a demo with Alice's sales team; the price depends on content volume, the abuse areas covered and whether intelligence and investigation services are included.
ActiveFence is a trust and safety service for platforms that host user posts, sold by Alice, the company formerly named ActiveFence. It combines AI detection and enforcement across more than 20 abuse areas with threat intelligence feeds, dark web monitoring and analyst investigations, drawing on adversarial data Alice says spans 120+ languages.
Automated detection and enforcement
Scores posts and accounts across Alice's full list of abuse areas at the scale of billions of items, with enforcement actions applied automatically.
ActiveMind moderation tools
SaaS moderation software powered by Alice's in-house language models that helps moderators validate, flag and respond to harmful content across text, image, video and audio.
Threat intelligence feeds
Live feeds surface harmful content, viral narratives and coordinated manipulation across languages and regions before they reach a platform's own users.
Deep and dark web monitoring
Round-the-clock monitoring of the open, deep and dark web finds bad actors and planned attacks aimed at a customer's platform.
Investigations and analyst teams
Global teams of analysts, researchers and policy experts trace networks and threat actors and advise on policy.
Adversarial indicator datasets
Verified datasets of domains, IPs, hashes and other indicators drawn from Alice's Rabbit Hole collection feed a platform's own detection.
| Tool | Price from | Free plan | Runs on | Rating |
|---|---|---|---|---|
| Quote | No | Web | 3.9 | |
| Quote | No | Web | 3.6 | |
| $169/mo | Yes | Web | 3.9 | |
| Quote | No | Web | 3.9 |
Proprietary
There is no public price list. Alice quotes ActiveFence after a demo, so the cost is only known once you describe your platform, your volumes and the services you want.
No free plan or self-serve trial is listed. The route in is a demo request through Alice's website.
Yes. ActiveFence renamed itself Alice in 2026. Alice's own site says ActiveFence is the former company name and the continuing name of its trust and safety offering for user-generated content.
Spectrum Labs, a contextual moderation vendor, was acquired by ActiveFence in September 2023, which said it would fold Spectrum's contextual text models into its platform (company announcement). Rival lists still describe the product as ActiveFence (Spectrum Labs).
Alice sells that separately as WonderFence, runtime guardrails that check prompts and responses in live AI apps and agents. ActiveFence itself covers content that people post on a platform.
ActiveFence is the user-generated content trust and safety offering of Alice, the New York and Tel Aviv company that was itself called ActiveFence until it renamed in 2026. The business started in 2018 protecting consumer platforms, and Alice's own site says ActiveFence remains the name of its UGC product while the company's newer work tests and guards AI models. SiliconANGLE reported a $140 million round led by Apax Digital Funds in 2026, taking total funding to $280 million.
It is more than a classifier API. ActiveFence combines automated detection and enforcement with ActiveMind, moderation software built on Alice's in-house language models, and with analysts who watch hidden forums and channels for attacks being planned against a platform. Behind it sits Rabbit Hole, the adversarial intelligence collection Alice says spans more than 120 languages, and teams of analysts, researchers and policy experts who run investigations for the customer.
It is built for the largest social networks, marketplaces, gaming and dating services, where harm is organised: extremist networks, child exploitation rings, coordinated manipulation and election interference. A head of trust and safety who needs intelligence on what is coming, not just a score on what was posted, is the buyer.
Most teams do not need that much. A product team that wants a self-serve API with published prices will get further with Deploy Content Safety, which publishes per-request rates, or a developer API such as Sightengine. A team whose problem is running its own review queues and appeals may want a case-management platform such as Cinder instead. Alice also sells WonderFence runtime guardrails for AI apps, which sit in Hokai's AI governance tools shelf and the AI governance and security guide rather than in content moderation.
The 2026 rename moved the company's headline story to AI security, and SiliconANGLE reports that its AI business grew more than 500% over two years. Buyers of the UGC product should ask how its roadmap and analyst staffing are protected inside a company now led by its AI work.
Product facts on this page (pricing, plans, integrations, compliance) are read from Alice (formerly ActiveFence)'s own pages and documentation. The verdict is Hokai's reading of them, not the vendor's claim.