Skip to main content

v4.14

Release Date: June 18, 2026

Highlights

  • Introducing Tribes: Tribes is a cohort-based analytical lens where Carbon Arc users can start with a brand, category, or company, and then ask questions from the point of view of a specific audience.
  • Docs & Library Tools: Carbon Arc's MCP Server is now able to answer questions about the analysis in our content library as well as surface more attributes (frequency, granularity, bias) from data assets available on the platform.

Key Releases

Introducing Tribes

Tribes is a cohort-based lens for audience analysis. Pick a entity (e.g. a brand like Walmart, a category, or a company) & a Tribe (such as Millennials), then ask what that audience actually does and what else they care about.

For example: among Millennial Walmart shoppers, which other brands show the strongest behavioral affinity?

Walmart x Millennial Brand Affinity in Tribes

The answer is not just "popular brands." It is the brands that over-index relative to what random chance would predict. That distinction matters because Tribes is designed to surface what is distinctive about an audience, not simply what is large, obvious, or assumed.

For marketers, investors, and operators, Tribes helps answer the same core question: who is this audience, what do they actually do, and what else do they care about?

There are a lot of ways to imagine or simulate how people might behave. We built Tribes to start with how they already do.

Use Tribes in Lenses today, or go deeper in Views for full rankings, trends, and comparisons. Read the Tribes docs →

Docs & Library tools (MCP)

Two new MCP tools let you look up platform and data information without context-switching.

Search Docs — Get answers about Carbon Arc products, workflows, and published analysis.
Try it: Ask about Iran-related content, then request the data behind the analysis.

Data Library — Explore what datasets exist, how they refresh, and what metadata is available.
Try it: Ask about Facteus bias or Clickstream refresh frequency.

Both tools are available in the MCP Server now. See Search Docs and Data Library for details.