Vexa

Vexa is an open-source meeting bot API that records, transcribes, and streams live audio data from Google Meet, Microsoft Teams, and Zoom meetings.

Vexa provides an infrastructure for capturing and processing live meeting data, including audio recording and real-time transcription. It functions as a set of programmable bots that join calls on platforms like Google Meet, Microsoft Teams, and Zoom. Developers can access meeting data through REST APIs, WebSockets, or an MCP server, allowing integration with AI agents like Claude Code, Cursor, or Codex. The system manages speaker-diarized transcripts and stores data in a Postgres database. Users can choose between a managed SaaS version or a self-hosted deployment on Kubernetes, OpenShift, or Docker, ensuring data sovereignty.

Vexa at a glance

Best for: Software engineer, Product manager
Runs on: Web, API, self-hosted
Pricing: Self-host for free; Individual $12/seat/month; Build on Vexa $0.30/bot-hour (plus $0.20/hour for transcription)
Released: February 2025
Category: AI audio tool
Website: vexa.ai

Key features

Real-time speaker-diarized transcription
WebSocket data streaming
Bi-directional conversational bot interface
Postgres-based storage for transcripts and audio
Open-source Apache 2.0 license
MCP server for AI agent connectivity
Self-hosted deployment options via Docker, Kubernetes, or OpenShift

Use it for

Connecting live meeting data to AI agents via MCP
Integrating meeting bots directly into internal software products
Automating meeting notes and data capture for teams
Archiving and querying historical meeting conversations

Alternatives to Vexa

AssemblyAI — Industry-leading speech models deliver high accuracy and low latency for voice applications.
Maybloc — Separates up to 25 speakers and preserves overlapping speech.
LiveKit Agents — Connects AI models to WebRTC rooms with reliable turn detection and streaming pipelines.
Vosk — It enables offline, low-latency speech recognition on small devices.
Pipecat — Builds real-time voice and multimodal agents with frame-level control.
Speech-to-Speech — Fully modular, swappable components with local or hosted execution options.
OpenWhispr — Run local speech-to-text models or use cloud processing with complete data privacy.
Vision Agents — It enables building low-latency AI agents with support for any model.
WhisperX — It enables 70x faster, batched transcription with high-accuracy alignment.
Speaches — It acts as an Ollama equivalent for running local TTS and STT models.
VoiceMem — It provides low-latency memory retrieval with emotional and personality context.
OpenLess — Keeps data local and allows the use of your own preferred ASR and LLM providers.

All Vexa alternatives