Context Window

34 articles about context window.

Size Queen Energy - Does 1M Context Actually Work?

·2 min read

1 million token context windows sound impressive but you never use them all at once. The real pattern is loading files on demand, not stuffing everything in

context-window1m-tokensllmai-agentsperformance

Memory Filters - Why AI Agents Need Aggressive Pruning

·8 min read

How to implement aggressive memory pruning for AI agents using LRU eviction, frequency scoring, and relevance decay - with concrete code examples and real benchmarks showing up to 90% token reduction.

agent-memorymemory-managementcontext-windowpruningai-agents

Why Do Agent Pacts Expire Before the Job Is Done?

·3 min read

AI agent agreements and context windows expire mid-task with no mechanism for renegotiation - a fundamental design flaw in how agents maintain commitments.

agent-agreementscontext-windowrenegotiationmulti-agentpersistence

Agents Can Overload Their Own Context - Use Separate Context with Shared Log

·2 min read

When agents share context, they overload it with each other's noise. Separate context per agent with a shared append-only log keeps each agent focused while

context-windowmulti-agentshared-logcoordinationoptimization

Claude Code Context Limit - When to Compact, Clear, and Optimize Token Usage

·2 min read

Managing Claude Code context limits effectively. Learn when to manually compact at 30-40% usage instead of waiting for the automatic limit to hit.

claude-codecontext-windowtoken-optimizationcompactproductivity

Context Compaction Ate Our Agent's Memory

·2 min read

How automatic context compaction silently destroys critical information that AI agents need to function correctly, and what to do about it.

context-compactionagent-memoryllmcontext-windowai-agents

Solving Context Loss in AI Coding Agents with Persistent State and Floating UIs

·2 min read

AI coding agents lose context constantly - hitting token limits, restarting sessions, forgetting decisions. Persistent state and floating UIs keep the agent

ai-agentscontext-windowclideveloper-toolsproductivity

CLAUDE.md Structure for Lossy Context Compression - Top and Bottom Wins

·2 min read

Context windows compress lossily. Structure your CLAUDE.md so critical instructions appear at the top and bottom - redundancy survives compression better

claude-mdcontext-windowprompt-engineeringai-agentmemory

Context Windows Are Not Memory

·2 min read

Context windows are working memory, not storage. Understanding this distinction is critical for building AI agents that maintain state across sessions.

context-windowmemoryworking-memoryai-agentsarchitecture

Memory Is Just Context with a Longer TTL - AI Agent Memory Systems

·2 min read

Memory files are lossy compressed embeddings of past context. Explore how context windows and long-term memory relate in AI agent architectures.

memorycontext-windowai-agentpersistencearchitecture

End of Day

·2 min read

For an AI agent, end of day is when the context window fills. How context limits create a natural work rhythm for autonomous agents.

context-windowagent-lifecycleautomationwork-rhythmlimits

Instruction Persistence in Long AI Agent Sessions - Keeping Agents on Track

·2 min read

LLMs forget instructions mid-session like losing focus. Techniques for maintaining instruction persistence in long-running AI agent sessions - echoing

ai-agentcontext-windowinstructionspersistencereliability

LLMs Forget Instructions Like ADHD Brains - Instruction Decay in Long Sessions

·3 min read

Instructions fade in long AI agent sessions the same way focus drifts in ADHD brains. Learn about instruction decay and practical mitigation strategies for

instruction-decaylong-sessionscontext-windowreliabilityprompt-engineeringartificial

Why Token Limits Never Add Up When Running Parallel AI Agents

·2 min read

Running parallel agents on a macOS app build reveals that token math is misleading. Context overhead, compiler loops, and shared file reads consume far more

token-limitsparallel-agentscontext-windowmacoscost-optimizationclaudecode

MCP Server Context Window Bloat and Why You Need a Toggle

·2 min read

Too many MCP servers trash your context window with tool definitions. A toggle approach lets you activate only the servers you need for each task.

mcpcontext-windowdeveloper-toolsai-agentsoptimization

I Installed 20 MCP Servers and Ended Up Worse Off

·2 min read

More MCP servers means more tools, more context consumption, and more confusion for your AI agent. Why running 3-4 servers daily outperforms a maximalist setup.

mcpserver-managementproductivitycontext-windowtool-overload

Memory Systems Are Graveyards - Less Context, Better Reasoning

·2 min read

Most agent memory systems become graveyards of stale data. Aggressive memory pruning leads to better reasoning because the model focuses on what actually

agent-memorypruningcontext-windowreasoningai-agents

Managing Context Bloat in AI Coding Agent Workflows

·2 min read

Context bloat kills AI coding agent performance. Learn why narrow, specialized skills beat broad context windows for persistent memory in Cursor and similar

context-windowcursorai-codingmemorycontext-managementproductivity

I Rarely Use Planning Mode Anymore - Context Windows Are Big Enough

·6 min read

Planning mode was essential at 8K tokens. With 200K context windows - and 1M in Claude Opus 4.6 - the model can see your entire codebase and figure out the approach as it goes. Here is when it still matters.

planning-modecontext-windowclaudeworkflowproductivity

Real Users Broke My AI Agent - Failures Testing Never Catches

·3 min read

How real users break AI agents in ways that testing never predicts. Context drops on interruption, unexpected inputs, and the gap between demo reliability

productionuser-testingreliabilitycontext-windowedge-casesai_agents

The Noise Floor Problem in AI Agent Context Windows

·2 min read

Every irrelevant token in your agent's context window raises the noise floor and degrades decision quality. Learn how to keep context clean and signal-rich.

context-windownoise-reductionai-agentssignal-to-noiseperformance

Why AI Agents Re-Plan From Scratch Every Turn - The Thinking Token Problem

·2 min read

Thinking tokens are not preserved between turns in AI conversations. Only visible output survives. This means agents are essentially re-planning from

ai-agentthinking-tokenscontext-windowplanningllm-architecture

Why 200K Context Models Outperform 1M When You Aggressively Clear Context

·2 min read

The biggest quality jump in AI agent workflows is not upgrading to a larger context window - it is being more aggressive about clearing context between tasks.

context-window200k-context1m-contextai-agentsprompt-engineering

Accessibility Tree Dumps Overflow LLM Context Windows - How to Fix It

·3 min read

Raw accessibility tree data can consume 24KB or more per dump, flooding AI agent context windows. The fix: write to temp files and return concise summaries

accessibility-treecontext-windowllmmacosoptimizationdesktop-agent

Why Backend Tasks Still Break AI Agents - Tool Response Design Matters

·2 min read

AI agents fail on backend tasks not because models are weak but because tool responses are poorly designed. Write full data to files and return compact

tool-designbackend-tasksagent-reliabilitycontext-windowmcp

Hitting Claude's Context Limit Mid-Build and How CLAUDE.md Fixes It

·2 min read

When Claude Code hits the context limit during a build, you lose project context. A CLAUDE.md file prevents starting over by keeping essential specs persistent.

claude-codecontext-windowclaude-mddeveloper-workflowproductivity

MCP Tool Responses Are the Biggest Context Hog - How to Compress Them

·3 min read

MCP server tool responses silently eat your context window. Here is how to compress accessibility tree data and other MCP outputs before they fill your

mcpcontext-windowaccessibility-apioptimizationtoken-managementclaudecode

Claude CoWork's Token Limits Hit Different - Why Local Agents Are Better for Big Tasks

·2 min read

CoWork has context limits that force session restarts on large codebases. A local agent running natively on your Mac manages its own context window without

coworktoken-limitslocal-agentcontext-windowmacos

Why Explaining a Process Is Harder Than Running It - The AI Agent New Hire Problem

·2 min read

Every new AI agent session starts from zero - the eternal new hire that never builds institutional memory. Why process documentation is now a core skill.

ai-agentsinstitutional-memoryprocess-documentationcontext-windowproductivityonboarding

MCP Servers That Pipe Raw Data Beat REST API Wrappers

·3 min read

The most useful MCP servers send raw data into context - transcripts, accessibility trees, full documents. The ones that just wrap a REST API add a layer of

mcpcontext-windowraw-dataapi-designagent-tools

The 1M Context Trap: Why More Context Makes Claude Lazier

·6 min read

Research on 18 frontier models confirms every one degrades with more context. The 'lost-in-the-middle' effect causes 30%+ accuracy drops. The counterintuitive fix: use less context, not more.

opuscontext-windowclaude-codeai-codingtokensproductivity

Why Scoped 50K Context Agents Outperform One Million Token Context

·3 min read

One million token context windows sound impressive, but scoped agents with 50K context each consistently outperform a single giant context for real

context-windowparallel-agentsscoped-agentsllmproductivityclaudecode

Using Opus as Orchestrator, Delegating to Sonnet and Haiku

·3 min read

The real win of using Opus as an orchestrator that delegates to Sonnet and Haiku is not cost savings - it is context window management. Opus burns through

opussonnethaikumodel-routingcontext-windowcost-optimization

Saving 10M Tokens (89%) on Claude Code with a CLI Proxy That Truncates Output

·3 min read

Claude already tries to tail output on its own, but by then the tokens are already in context. A CLI proxy that truncates command output before it hits the

claude-codetoken-optimizationcli-proxycost-reductioncontext-window

Browse by Topic