These updates focus on putting data and AI to practical use: from exploring the state of the Internet and transcribing meetings to developing software and editing video. ElevenLabs is also expanding its listening service in Brazil and its presence in Southeast Asia, with applications tailored to local languages and needs.
Cloudflare redesigns Radar around a map
Cloudflare has redesigned Radar to make Internet data accessible to a broader range of users while preserving depth and rigor. Launched in 2020, Radar provides a real-time view of global Internet trends based on Cloudflare’s network and has attracted network operators and academic researchers. The new landing page centers on a map, prioritizing Internet traffic and disruptions at the top. The map also guides viewers from the global picture to individual countries, trends and data stories. Below it, chart tiles have been replaced with a tabbed interface featuring summary statistics and paths to deeper exploration. This layout provides a clearer starting point for nontechnical users.
Source: Cloudflare
ElevenLabs explains how to integrate transcription into meeting workflows
ElevenLabs has published a guide to building meeting transcription tools with Scribe v2 and Scribe v2 Realtime, clarifying the choice between live and batch processing. The API converts audio into transcripts with timestamps and speaker labels. Real-time transcription generates low-latency captions during meetings, while batch processing uses recordings after meetings have ended. According to ElevenLabs, the models support more than 90 languages; the company also offers Zero Retention Mode for enterprise users. Developers can add context to help recognize terminology specific to a company, product or industry. The guide also covers connections to internal databases, CRM and enterprise software, helping engineering teams choose an appropriate integration approach.
Source: ElevenLabs
LegalOn uses model tiers to control AI costs
LegalOn has integrated Codex into its development workflow and expanded its use across the organization, shifting from choosing the most powerful model for every task to starting with a lightweight model and upgrading when needed. Under its internal criteria, Luna handles code implementation with clearly defined requirements; Sol supports routine design, data analysis and document drafting; and Astra is reserved for complex decisions such as architecture design. Published results report a 65% reduction in estimated daily costs and a 20% reduction in costs across mature business areas, while maintaining development speed. The company also sets monthly limits for departments and individuals and restricts Fast mode by default. AID CoE monitors and adjusts those limits based on business needs.
Source: OpenAI
Video Editor creates vertical videos with approval before cutting
The Grok Bot Marketplace page introduces Video Editor, created by Ethan Ng and described as a bot that edits raw footage into short, vertical talking-head videos. Its listed tasks include selecting the best takes, creating captions that match the actual speech, and adding B-roll, motion graphics and music. A notable step in the workflow is that the bot sends an editing plan for user approval before cutting the video. According to the description, the bot also checks each cut for synchronization and errors. The tool does not post or publish content on its own. This scope brings several post-production stages into one workflow while keeping plan approval with the user and separating publishing from the bot’s activities.
Source: x.ai
ElevenReader arrives in Brazil with Fábio Porchat’s voice
ElevenReader has launched in Brazil, allowing users to listen to written content in the voice of actor and comedian Fábio Porchat. As part of the launch, ElevenLabs produced an edition of Machado de Assis’s Dom Casmurro narrated by Porchat, with an original score and sound design. The app is available free on iOS and Android in Brazil; Dom Casmurro is free for all users, while the full licensed catalog is available through ElevenReader Ultra. Beyond the catalog, users can import ebooks, PDFs, articles and personal documents to listen to in the app’s voices, including Porchat’s. The launch combines literary content with the ability to turn personal documents into listening experiences, rather than limiting users to the books already available.
Source: ElevenLabs
ElevenLabs opens a Singapore office as its Southeast Asia hub
ElevenLabs has announced the opening of an office in Singapore to establish a Southeast Asia hub and strengthen its presence in Asia-Pacific. The company is expanding its local team to work with customers on building AI suited to the region’s languages and contexts. The applications cited include Funding Societies using ElevenLabs’ conversational AI to qualify leads across five markets; staff become involved only after the AI screening stage. Atome uses ElevenLabs in workflows involving interactions with employees and customers. The new office adds a local point of coordination, as the company’s technology is already being used for both customer interactions and internal operations in the region.
Source: ElevenLabs
Quick summary
Radar is moving to map-centered exploration of Internet data, combined with tabbed statistics.
ElevenLabs’ transcription guide distinguishes between real-time and batch processing, while also covering terminology customization and connections to internal systems.
LegalOn combines model tiers with usage limits; Video Editor retains an approval step before editing and does not publish on its own.
ElevenLabs is expanding in two directions: a listening service featuring Fábio Porchat’s voice in Brazil and a local team based at its Singapore hub.