Messages API Adds On-Demand Compaction in Beta: Developers Decide When to Compact, Not the System

Compaction timing is no longer the system's call — developers now decide when history gets condensed.

Adrian Koch · September 18, 2026

The official Claude Platform release notes added an entry on September 14, 2026: the Messages API can now compact a conversation on the server side at a moment the developer chooses, currently open for testing under the compact-2026-09-04 beta header. This feature extends the Compaction API that's existed since February 2026, but explicitly hands control over the triggering moment from automatic system judgment back to the caller.How it actually worksPer the official documentation, a...

practice

Your Hook Says "Blocking Error" But the File Still Changed? PostToolUse and PreToolUse Don't Actually Block the Same Way

An error message on screen can only tell you Claude received a signal worth...

advanced

Enabled Auto Mode and Thought You Were Safe? Permission Modes and Sandbox Boundaries Are Two Different Layers — Conflating Them Is How Things Break

The sandbox doesn't care what the model chose to run — only what the process...

news

Claude Cowork and Chat Officially Merge, Launching Claude Docs and Claude Slides Alongside It

Bring a quick question, or hand over a report — even after you close your...

Glossary

View All →

Adaptive Thinking

A reasoning mode where Claude dynamically decides whether and how deeply to think based on request complexity, replacing the older extended thinking mode that required manually setting a fixed token budget.

What is adaptive thinking, and how does it differ from the older extended thinking? Adaptive thinking is Claude's current mainstream reasoning...

intermediate

Chain-of-Thought Prompting

Prompting the model to write out intermediate reasoning steps before giving a final answer, rather than jumping straight to the answer, to improve accuracy on complex reasoning tasks.

**What is <a href="https://claude-me.com/en/glossary/prompt-techniques/chain-of-thought/" target="_blank" rel="noopener">Chain-of-Thought...

beginner

Few-Shot Prompting

Instead of describing in words what output you want, giving a few concrete input-output example pairs directly, letting Claude infer the style and format you're after from the examples themselves — often conveying detail more precisely than a written description can.

What is <a href="/en/glossary/prompt-engineering/few-shot-prompting/">Few-Shot Prompting</a>, and how does it differ from purely describing the...

beginner

Weekly Picks

prompt-examples

How to Write Your First Custom Slash Command in Claude Code: A Working Example From Scratch

If a tutorial's example command still shows a /project: prefix, that tutorial has already fallen...

Ryan Castillo

skill-library

Claude Code Merged Custom Commands Into Skills — Most "Commands vs Skills" Guides Are Already Outdated

The question worth asking now isn't "do I want automatic or manual triggering" — it's "does this...

Grace Lindqvist

practice

Why Your Claude Code Skill Silently Stops Triggering: The 15,000-Character Description Budget No One Warns You About

You think your description isn't good enough. Claude never saw it this conversation.

Sienna Park

beginners

Effort vs. Temperature in Claude: What's the Difference, and Why One of Them No Longer Works on Newer Models

Temperature 0 can still give you a shallow answer — it governs word-selection randomness, not...

Adrian Koch