Visual debugger for AI agent developers. Inspect context segments, track token usage, and identify memory bottlenecks before they break your agent.
Everything you need to understand and optimize your agent's context window.
Color-coded breakdown of system prompts, tools, conversation history, memories, and RAG documents.
Real-time token counts with percentage of max window. Supports GPT-4, Claude, and custom models.
Automatic detection when segments approach limits, with actionable optimization suggestions.
Deep-dive into system prompts with length analysis, section breakdown, and compression hints.
Export full context snapshots for sharing, version control, or automated analysis.
Compare context usage across sessions to identify regressions and optimization wins.
Paste an API response payload or connect via API key
Context is parsed into color-coded segments with token counts
See which segments are consuming the most tokens
Export, modify, and re-test your optimized context
“Finally I can see why my agent loses context after 20 turns. Conversation history was eating 70% of my window.”
“ContextLens saved us hours of debugging. We found 3 duplicate memory entries that were wasting 8K tokens.”
“The segment visualizer is a game-changer. I optimized my tool definitions and got 40% more conversation room.”
Free for solo developers. No credit card required.
Open Debugger →