AI agents // 2026–
The Agent Times
A news publication for people to read and AI agents to query.
- Role
- Co-founder · Product; editor & publisher
- Company
- Acton Labs
- Contribution
- Product direction, access, and editorial experience
- Result
- Live publication and public MCP endpoint

The problem
How can the same reporting serve readers and AI agents? Each needs a different interface, with access to the story’s sources and uncertainty.
My contribution
I shaped the product, the way agents get access, and the editorial experience. Both audiences read the same reporting, and its sources and uncertainty stay visible.
Two ways to access the news
The web publication provides sections, story pages, and editorial context for human readers. A public MCP interface provides another way to access the news corpus.
Sources, confidence labels, and provenance records make the basis of a story easier to inspect. They are signals to evaluate, not a guarantee that every article is correct.
Team & constraints
The masthead documents the editorial team and my role.
The product has to balance fast publication with sourcing and corrections. Reader-facing explanations and structured agent access serve different uses of the same reporting; the editorial team still owns the judgment behind it.
Current status
The publication and its agent interface are live. The product’s dashboards report internal audience and benchmark data, which change over time.
Internal data Live benchmark dashboard. First-party benchmark, maintained by The Agent Times; not an independent evaluation.