Back to Changelog
/ v0.2.0

AI-Powered API Discovery

GlacierHub can now automatically discover and connect to thousands of public APIs using AI-driven search and documentation analysis.

feature ai api-discovery

What’s new

GlacierHub now includes a multi-strategy API discovery pipeline. When you ask for data we don’t already have a connector for, the system searches public registries, the web, and its own knowledge graph to find the right API automatically.

Highlights

  • Knowledge Graph caching — successful discoveries are stored and reused, making repeat queries near-instant.
  • Brave Search integration — the system searches the web for niche APIs when registries come up empty.
  • AI documentation analysis — once an API is found, the AI reads its docs and configures the endpoint without any manual setup.
  • Confidence scoring — every discovered API gets a confidence score that adjusts over time based on success and failure rates.

How it works

Your query → Knowledge Graph check → Registry search → Web search → AI doc analysis → Chart

Each tier only runs if the previous one didn’t find a match. Most common queries resolve from the knowledge graph in under a second.

Bug fixes

  • Fixed an issue where single-value API responses rendered as empty charts instead of KPI cards.
  • Resolved a race condition in the comparison query pipeline for dynamically-discovered APIs.