Webinar

Transform your subscription web funnel: The psychology of why people pay

Save your seat
Industry NewsAI

New models, multi-modal chat from AI text and speech leaders Lindy, Type, and Deepgram

Published

Mar 4, 2025

TopicAI
Credit: type.ai

TL;DR

  • Type.ai introduces Chat Attachments and upgrades to the Claude 3.7 Sonnet model for enhanced AI-assisted writing capabilities.
  • Lindy.ai expands its AI assistants with image upload support in Gemini, enabling more interactive and context-aware responses.
  • Deepgram launches benchmark-leading Nova-3 Medical, a new AI-powered speech-to-text model that improves clinical transcription accuracy.
Author

Paddle News Desk

With the launch of GPT-4.5, upcoming GPT-5, and major funding for Anthropic, every application-layer startup is racing to deliver the fastest and most efficient access to NLP infrastructure. These three AI leaders in the text and speech space just announced new features and models to up-level user productivity and stay on top of the crowded space.

Better models, better results: AI-first document editor and Grammarly competitor Type.ai, has introduced two major updates—Chat Attachments and the Claude 3.7 Sonnet model. Users can now attach files, URLs, and text snippets directly within its AI-assisted chat-based writing tool. Whether users are spinning up brand guidelines, client briefs, or personal writing samples, the new feature enhances the AI’s ability to generate tailored responses that "understand" style and tone on the fly. Alongside this, Type has upgraded its default AI model to Claude 3.7 Sonnet to unlock faster state-of-the-art reasoning and improved long-form generation.

Multi-modal context awareness: Lindy.ai has expanded its AI assistants' capabilities with support for image uploads in Gemini. The company's agents are designed to manage your daily tasks and act as an EA for busy professionals. Lindy integrates with existing tools in order to automate mundane workflows and unlock productivity. With the new release, users can now share images with their AI assistants, allowing for more interactive and context-aware responses.

Leading healthcare AI: NVIDIA and Y Combinator-backed Deepgram has unveiled Nova-3 Medical, its latest AI-powered medical speech-to-text model, designed to set a new standard in clinical transcription accuracy. Deepgram is an AI-driven speech understanding and transcription tool, and is the most-funded speech AI company at its stage. The new model is optimized for performance and scale and outperforms competitors with a massive improvement in Word Error Rate when capturing critical medical terminology. Nova-3 Medical allows devs to fine-tune outputs with up to 100 custom terms, and is HIPAA-compliant for sensitive patient data.