---
title: "Context window breakdown | Langdock Changelog"
description: "We are introducing a context window indicator directly in the chat input bar, so you can see exactly how much of a model's context is being used, and what is…"
url: "https://langdock.com/changelog/context-window-breakdown"
language: "en"
---

[Changelog](https://langdock.com/changelog)

Jun 1, 2026

# Context window breakdown

We are introducing a context window indicator directly in the chat input bar, so you can see exactly how much of a model's context is being used, and what is using it.

#### Why we built this

Every AI conversation has a capacity limit (called the context window), a fixed amount of information the model can work with at any one time. As conversations grow, older messages are automatically summarized to free up space. Until now, there was no easy way to see how full the context window was or what was actually filling it up. This feature gives you visibility into what's going on behind the scenes and how much of the context window each part of the conversation uses.

#### What you can see

The indicator sits in the chat input bar and shows your current context window usage, for example, 13.6k / 200.0k (7%). You can click it to open a detailed breakdown.

The breakdown groups context usage into the following categories:

- **System tools**: tools available to the model in this chat
- **System prompt**: core instructions for the model
- **Messages**: the visible conversation between you and the model
- **Tools:** the built-in or integration actions performed in the chat
- **Attachments/images**: files and images you have added to the chat
- **Skills**: active skill descriptions available to the model
- **Memory**: saved memories included in the chat
- **Autocompact summary**: a summary of earlier messages
- **Free space**: how much context remains available

Each category shows a token count (the unit AI models use to measure text), a percentage, and a position in a stacked visual bar.

You can find more information about this in our [documentation](https://docs.langdock.com/resources/context-window).

We would love to hear how this works for you 🙌

Improvements

- **Skill sharing notifications**: Users now receive an inbox notification when a skill is shared with them, making it easier to discover and access newly shared skills.
- **Extended thinking on mobile**: Mobile app users can now enable extended thinking for supported reasoning models directly from the model selector.
- **Longer agent system prompts**: Agents now support system prompts of up to 50,000 characters, allowing you to create more detailed and comprehensive agent instructions.
- **Native tool search for latest models**: Agents using the latest OpenAI (GPT-5.4+) and Anthropic (Sonnet/Opus 4+) models now benefit from improved tool discovery and selection through native tool search.
- **Feedback version tracking and filtering**: Agent editors can now see which agent version each piece of feedback was submitted on and filter the feedback analytics page by version.
- **Shareable feedback link**: Agent editors can now copy a direct link to an agent's feedback dialog from the chat header menu, making it easier to collect targeted feedback.
- **Automated feedback prompts**: Users who have had more than 5 conversations with an agent are now prompted with a feedback banner to help collect actionable input for the agent editors.
- **Expanded dictation language support**: Additional languages including Polish are now supported for speech dictation and audio transcription.
- **Auto-hiding quick-action pills**: The quick-action pills below the chat input now automatically hide while you type and reappear when you clear your input, keeping the chat interface clean and focused.

[OlderClaude Opus 4.8](https://langdock.com/changelog/claude-opus-48) [NewerScheduled Tasks](https://langdock.com/changelog/scheduled-tasks)
