Convai
Voice-first AI NPC SDKs for Unity and Unreal, from ex-NVIDIA devs
Overview
A conversational-AI platform for building 3D NPCs and virtual humans, bundling speech recognition, LLM reasoning, TTS, memory, scene perception and action into Unity, Unreal and WebGL SDKs. Founded by ex-NVIDIA engineers; a direct competitor to Inworld AI.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 13 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Basic character creation
- Limited monthly usage
- Higher usage limits
- Unity/Unreal/WebGL SDKs
- More characters and interactions
- Priority features
- Expanded usage tier
- Scaling support
- Top self-serve usage tier
- Custom limits and SLA
What you can produce with Convai
- Unity, Unreal Engine and WebGL SDKs
- End-to-end voice: speech recognition, LLM reasoning and TTS
- Per-character knowledge banks (RAG)
- Scene perception and in-world action execution
- Long-term character memory
- Free tier plus usage-based self-serve paid plans
ASEAN Perspective
Convai in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
Convai is the platform game developers reach for when they want talking NPCs working in-engine fast — several report going from nothing to a pathfinding, conversing character in an afternoon inside Unreal or Unity. The stack is complete: speech in and out, LLM reasoning, per-character knowledge banks, scene perception, memory and action execution, wrapped in maintained Unity, Unreal and WebGL SDKs. The recurring complaint is latency, because conversations hit Convai's servers rather than running locally; for many projects the occasional pause is tolerable, but it rules the tool out for anything needing snappy, offline dialogue. Response quality also swings with how well you configure prompts and knowledge, so budget QA time to stop NPCs wandering off-topic. Pricing scales from a free tier through $29/mo indie up to four-figure business plans, plus usage. A sensible default for studios that want production-ready conversational NPCs and can live with cloud dependency.
What people say
Among developers, Convai has become close to the default for drop-in AI NPCs, and the praise is specific. On G2 and in developer write-ups, users single out how quickly it gets a character talking: one recounts building "a talking, pathfinding NPC in an afternoon" with the Unreal plugin and reckoning it "saves us months of R&D"; another calls the knowledge-bank workflow — load a character's facts, then just speak to it in-engine — "incredible for iteration" with "surprisingly robust" role-play. Reviewers consistently rate the Unity setup as easy and the character-customization options as deep.
The technical footprint is broad: real-time voice (ASR plus TTS), LLM reasoning, long-term memory, scene perception, and in-world action, delivered through Unity, Unreal and WebGL SDKs (all maintained openly under the Conv-AI GitHub org). It earned an NVIDIA developer-blog spotlight and a favorable review from Tommy Thompson of AI and Games, and it's frequently positioned as the price/performance leader against Inworld for engine-native NPC infrastructure.
The consistent knock is latency. Because dialogue is processed on Convai's servers rather than locally, developers report occasional pauses — one frames it as "an acceptable trade-off" for their project while flagging it as a real factor. A 2026 long-form review went as far as titling itself around a "latency problem that won't go away," attributing it to the cloud-round-trip architecture rather than to weak underlying tech. Anyone needing instant, offline, or twitch-timing dialogue should test this hard before committing.
The second caveat is output consistency: NPC conversation quality varies with prompts, knowledge configuration and the runtime context the game feeds in, so teams need iteration and QA to avoid off-topic answers, repetition or tone drift — more design effort than a deterministic dialogue tree.
Pricing is transparent and usage-based on top of plan fees. The Creator module runs Free ($0), Indie Dev ($29/mo), Professional ($99/mo), Scale ($499/mo) and Business ($1,199/mo), with Enterprise custom; reviewers estimate a typical moderate-interaction game at roughly $50–200/month in usage, while chatty MMOs can reach thousands. The upside is you pay for actual use, keeping prototyping cheap. Net: the strongest engine-native option for conversational NPCs if cloud dependency and tuning effort are acceptable.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including Convai's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Convai unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Convai directly →
Spotted something out of date? Suggest an update →
More in Code & Dev Tools
