Gemini API Now Lets Developers Mix Built-in and Custom Tools in a Single Request
Google shipped a quiet but high-impact update to the Gemini API this week: developers can now mix built-in Google tools — Search, Maps, and others — with their own custom functions in a single API call. Previously, combining these required multi-step orchestration that added latency and engineering complexity. The change is a direct response to developer feedback and eliminates one of the more frustrating friction points when building agentic Gemini applications.
Alongside tool mixing, Google introduced context circulation — a mechanism that lets the output of one tool feed directly into another within the same request. The net effect is that multi-step workflows that once required custom glue code can now be handled natively by the API, with less round-tripping and fewer opportunities for state to go stale between calls.
For smaller development teams and independent builders, this lowers the bar meaningfully. Agentic AI applications have historically demanded orchestration frameworks and significant infrastructure know-how. Google's change doesn't eliminate complexity, but it moves a real chunk of it into the platform — exactly where it belongs for an ecosystem trying to widen its developer base beyond large enterprises.