Dimension scores are derived from public data and fields; weighted into the composite. Reference only.
Rounda is a Japanese data consulting company founded in 2023, focusing on "Data × Generative AI" implementation support. It covers data infrastructure, ETL/ELT, DWH, BI, LLM implementation, and training. Its product, RoundaChat, is designed for small team pilots prior to company-wide procurement, allowing users to use GPT, Claude, Gemini, and local/custom models side-by-side within the same environment.
The core of RoundaChat is not a single chatbot, but a multi-LLM evaluation and team governance platform. It supports a single input for parallel generation by multiple models, allowing on-screen comparison of response quality, speed, token count, per-query cost, and latency. It also supports multimodal inputs such as images, PDFs, audio, video, and Office files. On the team side, users can share prompts, history, and evaluations, and conduct voting via Slack integration, making it ideal for forming model selection reports and internal AI usage guidelines.
Currently, the trial version is free for up to 5 users with full feature access, but it is available for a limited time and requires an accompanying interview. The tentative project plan is priced at 30,000 JPY/month for up to 10 users, including priority support and additional features in the official release. The official website also mentions a 100,000 JPY/month generative AI implementation project, but details are not expanded upon. Consulting and infrastructure building services are likely quoted per project.
Pros include support for OpenAI, Anthropic, Google, and OpenAI-compatible local models, integrating quality, cost, and team evaluation into a single workflow. It also features organizational data isolation, permissions, TOTP MFA, API Key encryption, and file MIME validation, making it enterprise-pilot friendly. Limitations include the possibility that official pricing and features may still change, and public materials do not specify key procurement information such as SLA, compliance certifications, and data residency. There is also no mention of a Chinese interface or Chinese-language support.
It is suitable for information systems, DX/AI promotion, corporate planning, product development, and model evaluation teams in Japanese enterprises, especially for scenarios requiring "small-scale validation first, then deciding which LLM to adopt company-wide." The official website does not provide information regarding access from mainland China, and given its reliance on APIs from OpenAI, Anthropic, and Google, there is uncertainty about network and payment availability. Domestic alternatives to consider include Dify, Coze, AI platforms from various cloud providers, or multi-model solutions like OpenRouter and LibreChat.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on rounda.co.jp official site.
rounda.co.jp is an Japan AI Apps provider. TG4G tracks its product information, an overall rating of 7.0/10, and a China-accessibility score of Workable. Click "Visit Official Site" to reach rounda.co.jp directly.