AnthropicapiMessage Threads in the Messages API (Public Beta)
A Claude API option that lets long-running agents send only what's new each turn instead of resending the whole conversation, for developers building on Claude.
Details
- You add a thread parameter to the request. The first request sends the full conversation and starts a thread, and later requests send only new messages plus the ID of the previous response.
- Threads are held in memory, encrypted, and not written to disk. A thread expires 5 minutes after its last request, and the client then resends the full conversation to start a new one.
- It shortens the wait, not the bill: standard token rates apply and token usage stays the same, because Claude still reads the full conversation each turn.
- In Claude Code, the median wait before a response starts dropped about 35% on requests of 500K to 1M tokens, and the median request in long conversations went from about 800 KB to 1.8 KB.
- In public beta on the Claude Platform, Claude Platform on AWS, Claude in Microsoft Foundry, and Claude Models on Google Cloud (Anthropic managed). Existing rate limits apply.
Best for
Long-running agents with very large conversations, such as coding agents, computer use agents, and document- or image-heavy workflows.
When to use it
Your agent is slow to start responding because every turn resends a huge conversation history.
Who benefits
Developers and technical teams running custom Claude agents. Most Claude Code users already have it on by default.
Join the newsletter to get ahead with AI.
I share strategies, new modules, free guides, and tools to get you better at using AI in your work.
Leading a team? See team programs →