product update

Google Docs adds persistent instructions for Gemini, capped at 1,000 per account

TL;DR

Google Docs now allows users to set persistent instructions for Gemini that apply across all documents. The feature, available to Google AI Plus subscribers in the US, supports up to 1,000 active instructions per account for controlling tone, style, and formatting.

2 min read
0

Google Docs adds persistent instructions for Gemini, capped at 1,000 per account

Google deployed an update to Gemini in Google Docs that allows users to set persistent instructions that the AI assistant remembers across all projects. The feature went live today for Google AI Plus subscribers and above.

Users can now input instructions via the Gemini side panel that define how the assistant should behave in all documents. According to Google, examples include "Always respond in bullet points," "When summarizing Docs, always include 3 bullet points at the top," or "Use a concise and professional tone for all my documents."

Technical specifications

  • Instruction limit: 1,000 active instructions per Google Account
  • Availability: US only, English language
  • Access tier: Google AI Plus and higher (free Gemini users excluded)
  • Visibility: Active instructions appear in the "sources" page and in Google Docs settings under the Gemini personalization section

How it works

Persistent instructions eliminate the need to repeat formatting, tone, and style preferences with each new document or prompt. Each active instruction appears as a source indicator beneath Gemini's responses. Users can review and manage all active instructions through the Gemini personalization section in Google Docs settings.

The feature addresses a documented friction point where users had to repeatedly specify the same preferences across multiple documents and sessions.

What this means

This update signals Google's move toward stateful AI assistants that maintain user preferences across sessions—a pattern also seen in OpenAI's Custom Instructions and Anthropic's Projects feature. The 1,000-instruction cap is generous for individual users but suggests Google expects enterprise accounts to accumulate substantial instruction sets over time. The US-only, English-only launch indicates Google is testing the feature's impact on response quality and computational overhead before expanding availability. The exclusion of free-tier users follows Google's broader strategy of gating advanced Gemini features behind paid subscriptions.

Related Articles

product update

Google Gemini Live gains access to Memory and Connected Apps from past conversations

Google has updated Gemini Live to access past conversation history through Memory and Connected Apps. The feature, currently available in English in the US, allows the voice assistant to reference previous chats and information from YouTube, Workspace, Utilities, and image generation tools during conversations.

product update

Mistral AI adds Deep Research agent, voice mode with Voxtral model to Le Chat

Mistral AI has released a major update to Le Chat, adding a Deep Research agent that generates structured research reports, a new voice input model called Voxtral, and Projects for organizing conversations. The update also includes multilingual reasoning powered by Mistral's Magistral model.

product update

U.S. government orders Anthropic to halt exports of Mythos and Fable AI models, both now offline for one week

The White House ordered Anthropic to restrict exports of its Mythos and Fable AI models last Friday, citing national security concerns. Anthropic pulled both models offline within 90 minutes of the Commerce Department directive, marking the first major test of AI export controls.

product update

GitHub details Qubot, internal Copilot-powered data analytics agent for plain language queries

GitHub has released technical details on Qubot, an internal analytics agent powered by GitHub Copilot that enables employees to query company data using natural language. The agent represents GitHub's implementation of AI-assisted data analysis for internal operations.

Comments

Loading...