Conversational Agent Builder
Vapi review
Developer platform for building production voice AI agents
Pros
- No-markup pass-through pricing on LLM, STT, and TTS costs keeps vendor choice flexible
- Handles telephony, transcription, and voice synthesis orchestration in one API
- Popular with developers building custom voice agent products, not just off-the-shelf bots
- Pay-as-you-go option with no subscription for smaller projects
Cons
- The advertised $0.05/minute headline understates real total cost once STT/LLM/TTS are added
- Concurrent call limits on pay-as-you-go plans (10 concurrent) may require Enterprise for scale
- Requires more technical setup than turnkey voice agent products
Our take
Vapi is infrastructure for developers who want to build a voice AI product rather than buy one off the shelf, stitching together telephony, speech-to-text, an LLM of your choice, and text-to-speech into a single API and billing each component at cost. That transparency is valuable for teams optimizing unit economics, but it also means the widely quoted per-minute price is only the platform fee, not the full cost of running a production voice agent.
Best for
Developers building phone or voice-based AI agents who want infrastructure without building telephony from scratch
Pricing
Platform fee from $0.05/minute pay-as-you-go, plus pass-through costs for LLM, speech-to-text, and text-to-speech (real-world total often $0.07-$0.30+/minute); enterprise plans custom, often $40K-$70K/year
Visit Vapi →