# Lower token spend and reduce drift with Claude context management

> Opt-in support for Claude’s context editing helps automatically remove stale tool results from long conversations.

- Published: 2025-11-12
- Author: Agno Team
- Category: Changelog
- Canonical: https://agno-com-nine.vercel.app/articles/lower-token-spend-and-reduce-drift-with-claude-context-management
- Markdown: https://agno-com-nine.vercel.app/articles/lower-token-spend-and-reduce-drift-with-claude-context-management.md

Opt-in support for Claude’s context editing helps automatically remove stale tool results from long conversations. This keeps prompts lean, cuts token usage, and reduces context drift — improving cost, latency, and answer quality over extended sessions.

**Details**

- Auto-clears outdated tool outputs to keep conversation history focused
- Lowers token counts without manual prompt engineering or ad hoc pruning
- Improves responsiveness and maintains accuracy in long-running chats
- Optional and backward-compatible

**Who this is for:** Teams running long-lived assistants (support, ops, internal copilots) where conversation length and cost control are critical.