Today I released a small but important update for PeakD ...unfortunately you cannot see it 😅

Content on peakd.com is now much easier for AI agents and automated tools to read and navigate properly.

[IMAGE: https://files.peakd.com/file/peakd-hive/asgarth/23xKhyRcPVqPpmZzdtNBgDW4BXiYRDUYagWMhSKKBz53sinBn8KxVpc81jtDniXMH5woN.png]

PeakD is a single-page app with specific SEO optimizations. That is good for humans because the site can stay fast, interactive, and flexible. But for agents, crawlers, assistants, a single-page app creates a practical problem: if they fetch a page like a normal HTML document, they will not see the full page content.

So PeakD now has a read-only Agent API.

A direct path to public PeakD content

The new agent-friendly setup gives automated clients a cleaner way to access public Hive content shown on PeakD.

Instead of scraping the rendered app, agents and tools can use public endpoints to get json and markdown formatted content for:

• posts
• snaps
• comments
• post discussions (json only)

The goal is simple: provide a clean, structured and documented path to read public content in an agent friendly format.

If you want you can give it a try yourself:

• Markdown: https://peakd.com/hive-139531/@asgarth/peakd-updates-governance-improvements-signup-flow-cleanup-stronger-analytics-and-more.md

• JSON: https://peakd.com/hive-139531/@asgarth/peakd-updates-governance-improvements-signup-flow-cleanup-stronger-analytics-and-more.json

• Discussion JSON: https://peakd.com/hive-139531/@asgarth/peakd-updates-governance-improvements-signup-flow-cleanup-stronger-analytics-and-more.discussion.json

Why Markdown matters

One of the new endpoints returns posts as Markdown.

That is useful because Markdown is much easier for language models and content tools to process than a fully rendered web page. It keeps the actual writing closer to its source form, without navigation, layout, modals, scripts, and all the other things that make sense in a browser but only add noise for an agent.

For example, this should help tools that:

• summarize PeakD posts
• extract quotes or references
• index public content
• build reading lists
• compare discussions
• help authors review older work

And there is a nice side effect as well. Reading the post in markdown is way more token efficient than parsing a full HTML rendered page.

Built for discovery

This API is read-only. It is for public content access. It does not let agents vote, comment, follow, edit posts, transfer funds, change settings, or perform any authenticated action.

Those actions can still be automated, but that's not something that we must solve within PeakD. Interacting with the Hive blockchain is already pretty doable with an AI agent and I have already shared a skill and a companion tool exactly for that purpose in the past.

Why I care about this

The web is changing quickly.

More content is being discovered through AI assistants, search agents, custom automations, research tools, and personal workflows that are not just "open a browser tab and scroll." If PeakD wants Hive content to remain easy to find and understand in that environment, we need to expose public content in formats those tools can use responsibly.

This is also important for authors. If you publish on PeakD, your content should be readable by people, and also readable by the tools people increasingly use to find and work with information.

This release is one step in that direction.

Cool, but does it work?

Well, I was not 100% sure myself so I tested on a few of different tools and models using my previous post. In the next screenshot you can see 3 different models on OpenCode, Codex and pi.dev figuring out how to load the markdown content by themselves.

[IMAGE: https://files.peakd.com/file/peakd-hive/asgarth/Eo1vym5mB6nWDjo4sVhHjepct9dqVWjivXehQainRQTGs6qkj1C2kuyYN9MLfLyvSmF.png]

Feedback welcome

As usual, let me know if you notice anything odd. Or if you use agents and you actually see a difference in your day to day experience.