Vapi

Voice AI infrastructure for developers

Developer Infrastructure Phone Voice Agents

Video & Audio Paid Has API

Researched 20 May 2026, 01:23 SGT · Published 19 May 2026, 01:23 SGT

Visit Vapi Compare alternatives

RECATOOLS Score

7.3 / 10

Capability

Value for money

Ease of use

ASEAN readiness

API quality

Founded

2023

San Francisco, California, USA

Users

—

Launched

—

Developer

—

Overview

Vapi provides voice-agent infrastructure for developers — build phone-call AI agents with one API. Handles speech-to-text, LLM orchestration, text-to-speech, telephony and call recording in a single platform. Used by customer-support automation, lead-qualification, scheduling and outbound-sales applications.

Use cases

Phone AI agents Customer support automation Lead qualification Appointment booking

What you can produce with Vapi

Build an inbound phone agent that answers calls, handles FAQs and routes callers, configured through the dashboard or API.
Run outbound calling campaigns for lead qualification or appointment reminders, triggered programmatically via the API.
Connect agents to calendars, CRMs and internal systems with tool calls and webhooks so a caller can actually book a slot or update a record mid-call.
Mix and match providers per agent — for example Deepgram transcription, a GPT or Claude model, and an ElevenLabs voice — without integrating each one separately.
Provision phone numbers or bring your own SIP trunk to put agents on real telephone lines.
Test agents by talking to them in the browser, then review call recordings, transcripts and analytics to iterate on prompts.
Orchestrate multi-agent squads that hand a call from one specialised agent to another mid-conversation.

ASEAN Perspective

Vapi 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).

RECATOOLS Verdict

Vapi is a developer-focused platform for building real-time voice AI agents, orchestrating speech-to-text, an LLM and text-to-speech into low-latency phone and web calls. Its appeal is flexibility: you bring your own model providers and telephony, with strong control over the pipeline, function calling and call logic. For engineers building custom voice assistants or call-center bots, it is one of the most capable building blocks available.

It suits technical teams comfortable wiring up STT/LLM/TTS providers, not non-coders wanting a turnkey bot. Honest caveats: the headline $0.05/min platform fee is misleading once you add third-party model and telephony costs, which push real all-in pricing toward $0.30/min; pay-as-you-go concurrency is capped; and it carries a real learning curve. The API and docs are its core product and are solid. Globally available and English-centric, with multilingual support dependent on the providers you plug in.

Independent AI-assisted assessment by RECATOOLS.

What people say

Vapi remains one of the most-used voice-agent infrastructure platforms, and its trajectory is strongly upward: a $50 million Series B led by Peak XV in May 2026 valued the company at $500 million, it reports over a billion calls handled and around a million developer accounts, and it won Amazon Ring's business over some 40 rival platforms. It is squarely an infrastructure play — speech-to-text, LLM orchestration, text-to-speech and telephony behind one API — rather than a finished call-centre product.

Developers praise the flexibility and speed of getting a working phone agent live. Being able to mix and match transcription, LLM and voice providers, wire in tool calls and webhooks, and test in the browser makes it a favourite for rapid prototyping, and the small G2 sample skews positive on setup experience.

The complaints are consistent and worth taking seriously. Latency is the big one: reviewers report response times that swing from under a second to four or five seconds, worsening at high concurrency or with heavier LLMs. Multiple users on Reddit and Trustpilot describe platform updates breaking previously working agents without warning, and support routed through a public Discord rather than dedicated account management; Vapi's Trustpilot score sat around 2.6/5 during our research, driven by support and stability friction. Costs also scale up quickly once usage reaches thousands of minutes, and several reviewers say the dashboard effectively requires a developer to understand.

Vapi fits engineering teams building custom voice agents who want maximum control over the stack and can absorb debugging and prompt-tuning work — agencies, SaaS teams adding voice features, and enterprises with in-house developers. Non-technical buyers wanting a turnkey AI receptionist will find the 80-percent-working reality frustrating and should look at packaged products built on top of platforms like this instead.

Summary of public user & expert reviews, compiled by RECATOOLS.

About this listing

Researched on Wednesday, 20 May 2026 at 01:23 SGT (UTC+8)

Published on Tuesday, 19 May 2026 at 01:23 SGT (UTC+8)

This entry was compiled from publicly available data including Vapi's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Vapi unless explicitly stated.

Data accuracy

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 Vapi directly →

Spotted something out of date? Suggest an update →