The new Intrascope

Intrascope positions itself as the “AI control layer of a company,” targeting enterprises that are struggling to manage their AI initiatives effectively.

A typical scenario is this: some people are using ChatGPT, others have purchased Claude on their own, while another team is testing Gemini. As a result, the billing details don’t match up, and the company’s documents and guidelines are scattered across various chat records. Intrascope brings together more than 120 models, including GPT, Claude, Gemini, Grok, DeepSeek, Qwen, Kimi, etc., into a single workbench. Administrators can specify which models each user is allowed to use and how much money they can spend on each model, and they can also keep track of every expense incurred.

Version 2.0 includes AI proxies, reusable skills, and intelligent model routing mechanisms. It also supports connections to over 1,650 applications, including MCP and RAG. Users can choose to use their own API keys or opt for the managed version, which starts at $59 per month. This version is ideal for teams of more than 20 people who are beginning to worry about the impact of “shadow AI”.