by Modulate
Modulate's real-time voice chat moderation for games and social apps, with plans from a free Starter tier to 200,000 voice hours a month.
Pick ToxMod if your game or social app has live voice chat and you need harassment and grooming caught as it happens, with audio evidence for moderators. Skip it if your harm is mostly text, images or video. Its free Starter tier covers 1,500 voice hours once, enough for a pilot.
It judges tone and escalation in live voice, not just a transcript.
It plugs into a game's own voice infrastructure and on-device pipeline (ToxMod product page, read 22 Nov 2026), while multimodal platforms that also handle audio compete for buyers who want one vendor.
Main trade-off
Voice only, with a steep jump from the free tier to the first paid tier.
| Plan | Monthly price | What it includes |
|---|---|---|
| Starter | Free | 1,500 voice hours, one-time, Additional hours 7.5 cents each |
| Epic | $5,000/mo | 50,000 voice hours a month, Additional hours 12 cents each, Automated escalation |
| Legendary | $10,000/mo | 100,000 voice hours a month, Additional hours 11 cents each, Custom category configuration, User report correlation |
| Mythic | $20,000/mo | 200,000 voice hours a month, Additional hours 10 cents each, Data insight support, Marketing and community consultation |
| Custom | Custom | As many voice hours as needed, Single-tenant support, Server-side SDK, Custom feature development priority |
Starter is $0 with 1,500 voice hours, granted once. Epic is $5,000 a month for 50,000 voice hours, Legendary $10,000 for 100,000 and Mythic $20,000 for 200,000. Extra hours are listed at 7.5 cents on Starter and 12, 11 and 10 cents on Epic, Legendary and Mythic. A Custom tier is quoted.
ToxMod is a real-time voice chat moderation service from Modulate for games and social platforms. It flags harassment, hate, threats and grooming in live voice by reading tone and escalation as well as words, and sends moderators the clip with context. Live voice data is deleted after 35 days by default.
Real-time voice monitoring
Analyses live voice channels and raises events for harassment, hate, threats, grooming indicators and coordinated abuse while the conversation is still going.
Tone-aware detection
Reads tone, intensity, escalation and interaction patterns as well as words, to separate real abuse from banter, sarcasm and baiting.
On-device first pass
Much of the analysis runs on the player's device and only clips likely to be harmful are sent on, unless a customer chooses to send all audio.
Moderator evidence
Each high-confidence event reaches moderators with the audio clip, timestamps and behavioural context for review.
Flexible enforcement
Signals can trigger in-game warnings or automated actions, or be routed into existing trust and safety tools and case management.
Custom categories and report correlation
Higher tiers add custom category configuration and matching of ToxMod events against player reports.
| Tool | Price from | Free plan | Runs on | Rating |
|---|---|---|---|---|
| $5,000/mo | Yes | Web | 3.6 | |
| Quote | No | Web | 3.8 | |
| Quote | No | Web | 3.6 | |
| Quote | No | Web | 3.9 |
Proprietary
Paid plans run from $5,000 to $20,000 a month for 50,000 to 200,000 included voice hours, with extra hours billed by the hour. Bigger platforms get a custom quote. The free Starter tier is a one-off allowance rather than a monthly plan.
There is a $0 Starter tier with 1,500 voice hours granted once, not every month. It suits a pilot or closed beta; continuous use needs a paid tier.
Harassment, hate, threats, grooming and exploitation indicators, escalation and conflict, and coordinated or repeated toxic behaviour in live voice chat. It also looks for contextual misuse such as sarcasm, baiting and dog whistles.
Not by default. Modulate screens audio on the player's device first and forwards only suspect clips, though a customer can choose to send everything. Benign conversations are not accessible to anyone and are deleted with the rest of the live data.
Activision is shown among the customers on Modulate's ToxMod page. Modulate does not publish a full customer list on that page.
ToxMod is the voice moderation product of Modulate, a Boston company whose wider business now sells voice intelligence for fraud detection, deepfake detection and contact centres. ToxMod is the part built for games and social platforms: it listens to live voice chat and flags harassment, hate, threats, grooming and other harm as it happens. Activision appears among the customers on the ToxMod page, beside a quote from a game studio's chief technology officer about using it in multiplayer voice.
The difference from text-first moderation is that it judges how something was said, not only the words. Modulate says ToxMod runs on Velma, which it calls an Ensemble Listening Model, reading tone, intensity, escalation and the dynamics between speakers, so sarcasm, baiting and dog whistles can be told apart from friendly trash talk. Where possible the first pass runs on the player's own device and only clips likely to contain harm go to Modulate's servers. Moderators get the flagged clip with timestamps and context.
It fits trust and safety and community teams at multiplayer games, social audio apps and voice-first communities that already have voice chat and need it covered in real time. Modulate also partners with Checkstep, which brings ToxMod's voice signals into the same review queue as text, image and video moderation.
It does nothing for written posts, images or video. A platform whose harm is mostly typed or uploaded should start with Sightengine, Deploy Content Safety or the free OpenAI Moderation API, none of which lists live voice chat moderation by tone the way ToxMod does. Modulate's fraud and contact-centre products are a separate buying decision; ToxMod sits with the other moderation tools in AI governance tools.
Product facts on this page (pricing, plans, integrations, compliance) are read from Modulate's own pages and documentation. The verdict is Hokai's reading of them, not the vendor's claim.