EarthLink launches Prompt Flow Studio for multi-model AI work

EarthLink Network Co., Ltd. is launching Prompt Flow Studio worldwide on the web Sept. 4, 2026, with iOS and Android coming next. The workspace brings together 400-plus AI models and adds enterprise controls for data protection, audit logs and governance starting in October.

Why it matters: - Prompt Flow Studio is designed to help teams compare, challenge and co-create across multiple AI models instead of relying on a single system. - The launch targets enterprise users that need data protection, auditability and control over which AI services can be used. - The product also makes model costs visible in real time, which can matter for teams managing AI budgets.

What happened: - EarthLink Network Co., Ltd. announced the worldwide launch of Prompt Flow Studio, including the United States and Japan, on Sept. 4, 2026. - The web version is available starting Sept. 4, 2026. - The iOS and Android apps will roll out on the App Store and Google Play next. - The product is available at Prompt Flow Studio.

The details: - Prompt Flow Studio lets users choose the right model for a task, compare answers side by side, run model-to-model debates and co-write documents with multiple models. - Uploaded internal documents can ground answers for any AI model the user chooses. - Each run shows its cost immediately. - The platform supports more than 50 AI providers and over 400 models across text, image, audio and video. - The model gallery shows each model’s strengths and cost. - An intent advisor suggests a model based on the task. - Model handover carries full context, including objectives, constraints and conclusions, when a user switches models mid-conversation. - The free plan includes Chat, Compare for text and image, Debate, Branch, Co-Draft, workflow automation in beta, model-agnostic RAG with 1 GB, personas, model handover and web search on lightweight models. - New sign-ups receive 300 PFC credits. - Paid plans unlock all models. - Enterprise data protection rolls out beginning in October 2026. - Organizations will be able to block sensitive data before transmission or mask it before sending and restore it after the response returns. - The rollout includes support for custom detection rules. - The platform adds tamper-evident audit logging with hash-chained records. - The enterprise controls also include Legal Hold, retention of up to seven years, selective deletion of personal data and governance over which AI services, models and standard personas an organization permits. - Conversations are private by design, and administrators do not get day-to-day access to users’ prompts or AI responses. - Legal requests will go through a controlled, fully audited disclosure process. - Prompt Flow Studio supports debate workflows that assign different models and personas to roles. - An advocate and critic exchange claims and rebuttals, a facilitator organizes the issues, a dedicated Red Team AI attacks weaknesses and a synthesizer produces the final conclusion with multiple quality metrics. - Prompt Flow Studio supports personas such as a strict auditor, a hands-on practitioner and a brand editor. - Organizations will be able to distribute standard personas across whole teams. - The co-draft workflow uses multiple AIs to propose outlines, review each other’s work and produce successive versions of a single document. - The model-agnostic RAG workflow grounds answers in uploaded documents and can be used in strict documents-only mode or blended mode. - The same grounded context can be answered by OpenAI, Anthropic, Google and other models. - Compare mode sends one prompt to several models at once and shows the results side by side. - Image comparison can generate results from multiple image models at the same time in a single grid. - Per-model costs are displayed for comparison workflows. - Every response shows confirmed cost in PFC credits, with line-item breakdowns for AI responses, knowledge retrieval and web search. - Monthly usage is tracked by mode, AI service and day. - An estimate appears before a user sends a prompt.

Between the lines: - EarthLink is positioning Prompt Flow Studio as an AI control layer, not just a chat app. - The emphasis on debate, red-teaming and multi-model comparison suggests a push toward more defensible outputs for enterprise workflows. - The privacy and audit features are likely aimed at organizations that need to adopt AI without giving administrators broad access to employee conversations. - The October enterprise rollout creates a staged launch, with broad access first and governance features following later.

What's next: - The iOS and Android apps are expected to arrive after the web launch. - Enterprise protection features will begin rolling out in October 2026. - EarthLink is directing users to Prompt Flow Studio for details and sign-up. - Media inquiries are directed to Prompt Flow Studio contact.

The bottom line: - Prompt Flow Studio is entering the market as a multi-model AI workspace built around comparison, collaboration, cost visibility and enterprise-grade controls.