Google previews Tool Retrieval in Gemini Interactions API to load tool schemas only when needed
Google announced Tool Retrieval on November 19, 2026, available in preview in the Gemini Interactions API. Agents can keep large tool libraries on standby and load full parameter schemas only when the current task needs them.
Google announced Tool Retrieval on November 19, 2026, making the feature available in preview in the Gemini Interactions API. It lets agents keep large tool libraries on standby and load full parameter schemas only when the current task requires them, according to a feed summary of the announcement published by Unite.AI.
What Google says
Google says the conventional approach becomes unwieldy as agents connect to dozens or hundreds of tools. Sending every full schema with each request, the company says, inflates input-token counts, adds latency and reduces tool-selection accuracy as the candidate list grows.
Tool Retrieval is Google's answer to that problem. Instead of placing every tool's complete parameter schema in the model's context up front, the agent holds the library in reserve and pulls in a full schema only when the task at hand calls for it.
Availability
- Feature: Tool Retrieval
- Status: Preview
- Surface: Gemini Interactions API
- Announced: November 19, 2026
- Pricing: not yet disclosed
The summary does not give details on pricing for the feature, supported models, or benchmark results.
What this means
Agents that connect to many tools pay for tool definitions on every request when all schemas are sent in full. Google's stated concerns are token volume, latency and selection accuracy, and Tool Retrieval targets all three by deferring schema loading until a tool is needed.
The feature is in preview, so its behavior and availability may change before general release. Developers building agents on the Gemini Interactions API with large tool libraries are the intended users. Google's claims about the cost of the conventional approach are the company's own, and the effect of Tool Retrieval on token counts, latency and accuracy has not been quantified in the announcement as summarized here.
Related Articles
Google makes Genie 4 available in Gemini API and AI Studio with 720p, 24 fps interactive worlds
Google made Genie 4 available through the Gemini API and Google AI Studio on November 19, 2026. The model generates continuous 720p video at 24 frames per second with synchronized native audio, and worlds respond to keyboard controls and live text prompts.
Google rolls out computer use in the Gemini app for Mac for multi-step desktop tasks
Google announced on November 19, 2026 that computer use is rolling out in the Gemini app for Mac. Gemini can carry out multi-step work across local files, desktop applications and websites, working in an isolated background window with a live preview in the chat.
Google builds isolated 'Mirror Dimension' web to train and test Chrome AI agents
Google's Chrome team built the Mirror Dimension, an isolated web ecosystem it fully controls, to train and test Gemini-based browsing agents without exposing them to the open internet. Google is detailing the project in a blog post, according to Wired.
Tunix agentic GRPO lifts Gemma 4 E2B from 9.4% to 95% win rate in Plane Strike
Google's developer blog details how Tunix's agentic GRPO pipeline trained Gemma 4 E2B to play Plane Strike with tool calls. The win rate on held-out boards rose from 9.4% to 95% after 100 steps on a single Cloud TPU v6e-8 host.
Comments
Loading...