Launches, new modules and the thinking behind Raq.com.
Page 4 of 9
Stop trusting someone else's benchmark. Write the prompts you actually care about, run them across the models you use, rate the responses by hand, and publish a sortable leaderboard at your own URL. Every step works manually or hands-off via Supercomputer and your own AI assistant.
Dual AI Chat compared two models side by side. Multi AI Chat compares up to ten, switches between answers in tabs, lets you carry more than one response into the next turn, runs web research and Python analysis, and writes a consensus summary across the responses.
The SEO Checker tool now audits for SEO and answer-engine readiness, and tracks whether AI answers mention, cite, or misrepresent your brand. See brand mentions, owned citations, competitor visibility, and accuracy issues over time.
Orchestrator lets an MCP-connected assistant control multiple local agent sessions. Tell one chat to launch or attach agents, queue prompts, chain work, continue when conditions are met, and summarise what needs attention.
Paste your homepage URL and Raq.com builds a shared brand workspace from it: logos, colours, fonts, voice. Every field is exposed via MCP, so Supercomputer, Claude, Codex, or any agent you work with can read your brand and keep its output on-brand without you re-explaining it every time.
GPT Image 2 launched straight to the top of the image leaderboards, and it is live in AI Image Lab today. Pixel-perfect text rendering, photorealistic lighting, and brand-accurate product shots, all running alongside the other models in the same batch workflow.
Name the companies and topics you care about. Raq.com gathers news and Companies House filings on a schedule, clusters the signal, and writes briefings you can act on. Add and query everything from Supercomputer, Claude, or any MCP-connected assistant.
Image Optimiser has grown beyond compression and format conversion. It now handles AI upscaling, object removal, watermark removal, photo restoration, black-and-white colourising, AI file naming, HEIC conversion, resize modes, and text watermarks, all from the same batch workflow.
Create one-time links and intake links for passwords, API keys, and other sensitive handoffs. Secure Send works from the web app, Supercomputer, and any MCP-connected assistant.
Describe a UI, tick the models you want, and watch them all generate in parallel. Rate and compare side by side, then download the winner as HTML or a PNG.
Convert documents, images, audio, and video in batch. Up to 50 files at a time, per-item progress tracking, private time-limited storage, and no need to trust a random site on the internet with your files.
Install a tiny agent on your Linux servers to track load, CPU, memory, disk, and processes. Or point cron jobs and scripts at a heartbeat URL. Every monitor is fully manageable from AI assistants via MCP.
We use optional analytics cookies to understand how people find and use this website.
