Lovable adds GPT-Live, Claude, Gemini 4 Argon and more for building AI apps

Lovable has added a set of AI models and tools that users can build into their apps, including GPT-Live from OpenAI, the latest Claude models from Anthropic, and Gemini 4 Argon, Gemini Omni and Lyria from Google. Tools from Perplexity, Firecrawl, Parallel and Jev by Typesafe AI are also included, according to the company.
- Lovable now offers GPT-Live, the latest Claude models, Gemini 4 Argon, Gemini Omni and Lyria inside its app builder.
- Perplexity, Firecrawl, Parallel and Jev by Typesafe AI add web search, scraping, research and classification.
- Lovable can pick a model automatically, or users can name one in the chat, such as "use Perplexity."
- Tools can be combined, for example Jev sorting support tickets while Claude drafts replies.
Voice, reasoning and video models
According to Lovable, GPT-Live from OpenAI lets apps offer natural, real-time conversation. Users can build voice assistants that help people find answers, navigate features and complete tasks, from booking an appointment to getting customer support. Lovable suggests a customer assistant that books appointments, a language tutor for practicing speaking before an exam, and a sales pitch coach that asks follow-up questions. One team member also built an app for talking to Rocky from Project Hail Mary.
The latest Claude models from Anthropic are aimed at jobs that need deep reasoning, careful planning and polished output. Lovable lists a tool that turns client briefs into finished proposals, an app that estimates the worth of an item from a photo of a marketplace listing, and an inbox assistant that flags urgent messages and drafts replies.
From Google, Gemini Omni is described as the company's latest video model that can animate photos or create videos from any input. Lyria writes original tracks with lyrics, in a style, tempo and vocals the user chooses. Gemini 4 Argon is suggested for supporting knowledge work. Lovable's examples include a shop tool that turns one product photo into a promo video with its own music, a real estate app that makes walkthrough clips from listing photos, and a market research app that reads competitor reports and trend charts and writes a campaign brief.
Web data and fast decision tools
Several of the additions focus on getting information from the web or making quick classifications.
- Perplexity gives apps real-time answers grounded in the live web, with citations people can check. Lovable suggests a morning briefing for investment teams or a city guide that answers "what's on this weekend?" with event links.
- Firecrawl is described as the knowledge library for AI. It lets an app search, scrape and pull clean data from anywhere, for example in a price tracker that watches competitor pages or a help-center assistant that answers from a company's own docs.
- Parallel is an AI web search platform that gathers and synthesizes information, from quick facts about yesterday's news to deep research such as mapping a company's investments over decades. Suggested uses include lead enrichment, a research paper monitor, a fundraise tracker and a grants and RFP finder.
- Jev by Typesafe AI makes quick decisions an app can act on, such as a category, a score, or a yes or no. It could tag live chat messages as a refund, a delivery question or urgent, or route support tickets to the right team.
Lovable also says the tools can be combined. Paired with Claude, Jev sorts incoming requests while Claude writes the reply.
How to use them and what it means
Lovable says users do not need an AI strategy. They describe the app or feature they want, for example "Add a research panel that gives me a cited summary of my competitors." Lovable will choose a model or tool, or the user can name one in the chat, such as "use Opus 5.5" or "use Perplexity."
For people building on the platform, the change means voice, document reasoning, video and music generation, live web search and ticket routing are available in one place, without leaving Lovable. The source does not give pricing, usage limits or availability details for the individual models and tools.
Lovable: The AI you already use, now powering the apps you create


