LIVE
Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|
Toolofficial publisher

Vapi

Tool : A precise action the AI can call · requires vapi

Full-featured voice agent with telephony and function calling.

The most complete solution for a telephony agent.

Docs
Overall score
5.9 /10
Security: one caveat, see the card
Utility
7.0
Reliability
6.0
Security
One caveat
Ease
Technical · Technical

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

Building a voice agent capable of making and receiving real phone calls requires managing speech recognition, a language model, text-to-speech, and telephony infrastructure all at once. Vapi handles this entire stack so developers can focus on business logic.

In practice, each Vapi assistant combines three components: a speech-to-text transcriber, a large language model that generates responses, and a text-to-speech engine. All three are fully configurable across dozens of providers including OpenAI, Anthropic, Google, Deepgram, and ElevenLabs. Inbound and outbound calls go through phone numbers managed within the platform, and the agent can trigger external actions via tools and webhooks.

Vapi is aimed at developers who want to automate phone workflows: customer support, lead qualification, appointment scheduling, medical triage, or order management. The platform offers two building modes: Assistants (a single system prompt with tools) for most use cases, and Squads (multiple specialized assistants with context-preserving transfers) for complex scenarios.

However, Vapi is not suited for non-technical users: access is exclusively through an API, and advanced configuration requires familiarity with concepts such as webhooks, function calling, and third-party provider integrations.

What it offers

  • ·Inbound and outbound phone calls managed directly from the platform, with dedicated phone numbers.
  • ·Sub-600ms response times for natural real-time conversations.
  • ·Free choice of provider for each component: transcription, language model, and text-to-speech (OpenAI, Anthropic, Google, ElevenLabs, Deepgram, and more).
  • ·Web integration to embed live voice calls directly into an application.
  • ·Squads mode: orchestration of multiple specialized assistants with context-preserving transfers, suited for complex workflows.
  • ·Function calling and webhooks to connect the agent to APIs, databases, or existing systems.
  • ·Model Intelligence presets that configure transcriber, model, and voice in one click for common use cases.
  • ·Vapi CLI to manage the entire platform from the terminal.

In practice

Price
Paid
Commercial use
Commercial use: yes
Available on
api
Host
vapi
FR interface
No

Security and data

Rating
Security: one caveat, see the card
Capabilities
readswritesexecutesnetwork
Your data
Data with third parties
Maintenance
Active

Links and repository checked on 25 October 2026 · added on 1 June 2023

How to install Vapi

1. Create an account at vapi.ai and retrieve your API key from the dashboard. 2. Install the Vapi CLI using the provided command to manage the platform from your terminal. 3. Configure a phone number (free Vapi number or existing telephony integration) in the Phone Numbers section. 4. Create an Assistant by defining the system prompt, transcription provider, language model, and voice. 5. Add your tools and enter your webhook server URL for call events, enabling server authentication to secure incoming requests. 6. Test an inbound or outbound call from the dashboard before going to production.
curl -sSL https://vapi.ai/install.sh | bash

Official documentation ↗

In the same category

Whole category Automation & agents →