AIAny
Icon for item

Vexa

Runs a self-hosted meeting bot and transcription API that joins Google Meet, Teams and Zoom and streams speaker-attributed transcripts in real time. Compiles meetings into a git-backed Markdown workspace and runs sandboxed agents on your infrastructure; Apache-2.0 and air-gap capable.

Introduction

Most meeting-AI products rely on vendor clouds and retrofitted transcripts. Vexa flips that model: it captures the conversation inside your perimeter and turns meetings into a portable, git-managed knowledge workspace that agents can safely operate on. That upstream capture changes what you can automate and audit.

What Sets It Apart
  • Bot-in-call capture with speaker attribution — the bot actually joins Meet, Teams, Zoom (and Jitsi) and streams live, attributed transcripts, not just ingesting recordings after the fact. So what: you get real-time context for agents and live copilot scenarios instead of delayed summaries.
  • Knowledge-as-code output — every meeting compiles into Markdown files in a git workspace (OKF-style kg/ bundle). So what: history, diffs, and provenance are first-class; teams own and fork their meeting knowledge.
  • Sandboxed agent runtime — agents run as ephemeral, isolated workloads (Docker, process, or Kubernetes Pods) with configurable mounts and no egress by default. So what: you can automate post-meeting processing, briefs, and actions without exposing data to third-party inference services.
  • Designed for air-gapped and regulated environments — bring-your-own models or local STT; many deployment artifacts (CALM architecture, security manifests) are provided. So what: banks and healthcare orgs can run meeting intelligence without data leaving their network.
Who It's For and Trade-offs

Great fit if you need in‑perimeter meeting capture, want transcripts tied to versioned files, and must control model and network egress. It’s aimed at teams and enterprises that can operate Docker/Kubernetes and want long-term ownership of meeting data.

Look elsewhere if you only need a lightweight client-side notes app, cannot run containers or provide modest compute (the full stack expects at least ~8 vCPUs/16 GB for production make all), or if you prefer a vendor-managed, hosted SaaS with no operational burden.

Where It Fits

Positioned between hosted meeting-AI services (which process audio in vendor clouds) and DIY approaches (Whisper + ad-hoc bots). Vexa is the open, deployable middle: full capture + agentic automation that you run and control.

More Items

GitHub
AI Infra2025

Measures generative AI inference performance with token-level metrics (TTFT, inter-token latency), latency, and throughput under realistic traffic patterns. Provides a multiprocess engine, real-time TUI dashboard, extensible plugins, and integrations for telemetry and result uploads, aimed at inference benchmarking and capacity planning.

GitHub
AI Agent2026

A curated collection of production-ready Agent Skills that turn tasks—presentation production, image generation, local KB retrieval, article assembly, and web-design—into agent-loadable skill folders. Uses a SKILL.md contract, supports multiple agent runtimes (Claude Code, Cursor, Codex), and offers modular install paths with pinned release zips.

GitHub
AI Train2019

Train and experiment with multi-billion to trillion-parameter transformer models on large GPU clusters using GPU-optimized building blocks and reference training scripts; offers advanced parallelism and mixed-precision support for research teams and ML engineers.