Soniox

Tray.ai

Tray.ai is an AI-native enterprise integration platform as a service (iPaaS) that enables mid-market to enterprise teams to build AI agents, govern Model Context Protocol (MCP), and integrate over 700 apps. It combines low-code automation with developer-friendly features for orchestrating complex, data-heavy workflows. The platform distinguishes itself by unifying data integration, automation, MCP governance, and AI agent building in a single solution.

Bland AI

Bland is an enterprise voice AI platform that lets businesses build, deploy, and monitor AI phone agents capable of handling calls and hold waiting. It targets companies that cannot afford to miss calls but cannot hire more staff, offering self-hosted models, sub-second latency, and enterprise compliance. The platform is trusted by 250+ enterprises and claims to have added $40M in revenue in five months.

Rev

Rev provides transcription services combining AI automation with human review for high accuracy. It serves businesses and developers needing reliable captions or transcripts, offering a real-time streaming API for live captioning integration. Its distinctive value is the blend of speed from AI and precision from human editors, though it comes at a premium price.

Gladia

Gladia provides a real-time transcription and audio intelligence API designed for developers who need low-latency speech-to-text and audio analysis. It differentiates with a focus on developer experience and has raised $16M to scale its platform.

Regent

Regent is an AI monitoring platform that detects and alerts on behavioral changes in AI models. It serves engineering and ML teams who need to ensure their deployed AI systems remain reliable and predictable. Its key distinction is real-time detection of behavioral drift without requiring manual thresholds.

Grok Voice Think Fast 1.0

Grok Voice Think Fast 1.0 is an API-accessible voice agent from xAI, designed for developers to integrate advanced voice capabilities into their applications. It leverages xAI's most capable voice model, enabling real-time, natural speech interactions. The product targets AI builders seeking to add voice interfaces without building from scratch.