Vera Platform by Cortex Research

The platform runs on Vera 1.6, Cortex Research's own flagship model. It handles text, images, and video in one system, with a 1M-token context window that lets it work through long documents without losing the thread. Cortex Research built it specifically for agentic workloads, meaning multi-step tasks where the model has to plan, act, and check its own work.

Vera ships with a catalog of ready-made agents, each tuned for a job. There's an Analyst Agent for research, an HR Agent, an Email Agent, a Marketing Agent, a Financial Agent, a Coding Agent, an Invoicing Agent, a Sales Agent, a Compliance Agent, and a Legal Agent, among others. You pick the one that fits the task instead of prompting a general chatbot and hoping. Less guesswork. Faster results.

Agents connect to the apps your team already uses. Gmail, Slack, and Google Drive are live now, with more on the way. Once linked, Vera can read and write across those tools, so a request like "summarise this thread and draft a reply" actually touches your inbox rather than describing what you should do. It works inside your tools.

Vera browses the web as it works, cross-references multiple sources, and returns answers with citations. That's a meaningful difference from models that answer from training data alone, especially for anything time-sensitive like competitor moves or market news. No stale answers.

Both the models and the infrastructure are UK-built and UK-hosted. For organisations under UK data rules, this removes the usual workaround of anonymising data before sending it to an overseas provider. Cortex Research positions sovereignty as the core selling point of the platform. That's the whole pitch. Everything else follows from it.

Agents produce finished documents, reports, and slide decks, not just chat replies. You can feed in source material and get back something structured enough to share, which saves the step of pasting model output into a separate tool. No copy-paste shuffle.