Claude Blog 采集 (2026-05-17)¶
共采集 5 篇文章
📋 文章索引¶
- New capabilities for building agents on the Anthropic API - May 22, 2025 (评分: 9.5)
- Claude Sonnet 4 now supports 1M tokens of context - Aug 12, 2025 (评分: 9.0)
- Agent view in Claude Code - May 11, 2026 (评分: 9.0)
- Claude on Amazon Bedrock now available to every AWS customer - Sep 28, 2023 (评分: 9.0)
- Introducing the analysis tool in Claude.ai - Oct 24, 2024 (评分: 9.0)
New capabilities for building agents on the Anthropic API¶
来源: Claude Blog 发布日期: May 22, 2025 采集时间: 2026-05-17 价值评分: 9.5/10 正文字数: ~6162 字符
摘要¶
Claude now offers code execution, MCP server connections, file storage, and extended prompt caching through the API—giving developers powerful tools to build agents that analyze data, connect to external systems, and maintain context for longer periods of time.
正文内容¶
Claude now offers code execution, MCP server connections, file storage, and extended prompt caching through the API—giving developers powerful tools to build agents that analyze data, connect to external systems, and maintain context for longer periods of time.
Today, we're announcing four new capabilities on the Anthropic API that enable developers to build more powerful AI agents: the code execution tool, MCP connector, Files API, and the ability to cache prompts for up to one hour. Building better AI agents Together with Claude Opus 4 and Sonnet 4 , these beta features enable developers to build agents that execute code for advanced data analysis, connect to external systems through MCP servers, store and access files efficiently across sessions, and maintain context for up to 60 minutes with cost-effective caching—without building custom infrastructure. For example, a project management AI agent can use the MCP connector with Asana to reference tasks and assign work, upload relevant reports via the Files API, analyze progress and risks with the code execution tool, and maintain full context throughout—all while keeping costs down through extended prompt caching. These capabilities join existing features like web search and citations as part of a comprehensive toolkit for building AI agents. Read on to explore each new capability in detail. Code execution tool We're introducing a code execution tool on the Anthropic API, giving Claude the ability to run Python code in a sandboxed environment to produce computational results and data visualizations. This transforms Claude from a code-writing assistant into a data analyst that can iterate on visualizations, clean datasets, and derive insights directly within API calls. With the code execution tool, Claude can load datasets, generate exploratory charts, identify patterns, and iteratively refine outputs based on execution results—all within a single interaction. This means that Claude can handle complex analytical tasks end-to-end, rather than just suggesting code for you to run separately. Key use cases include: Financial modeling : Generate financial projections, analyze investment portfolios, and calculate complex financial metrics. Scientific computing : Execute simulations, process experimental data, and analyze research datasets. Business intelligence : Create automated reports, analyze sales data, and generate performance dashboards. Document processing : Extract and transform data across formats, generate formatted reports, and automate document workflows. Statistical analysis : Perform regression analysis, hypothesis testing, and predictive modeling on datasets. Organizations receive 50 free hours of usage with the code execution tool per day, then pay $0.05 per hour per container for additional usage. Explore the documentation to learn more about pricing. MCP connector The MCP connector on the Anthropic API enables developers to connect Claude to any remote Model Context Protocol (MCP) server without writing client code. Previously, connecting to MCP servers required building your own client harness to handle MCP connections. Now, the Anthropic API handles all connection management, tool discovery, and error handling automatically. Simply add a remote MCP server URL to your API request and you can immediately access powerful third-party tools, dramatically reducing the complexity of building tool-enabled agents. When Claude receives a request with MCP servers configured, it automatically: Connects to the specified MCP servers Retrieves available tools Reasons about what tool to call and what arguments to pass Executes tool calls agentically until a sufficient result is achieved Manages authentication and error handling Returns the enhanced response with integrated data The growing ecosystem of remote MCP servers means you can easily add capabilities to your AI applications without building one-off integrations. You can integrate with any remote MCP server, including those from Zapier and Asana . See more remote MCP servers in our documentation . Files API The Files API simplifies how developers store and access documents when building with Claude. Instead of managing file uploads in every request, you can now upload documents once and reference them repeatedly across conversations. This streamlines development workflows, particularly for applications that need to work with large document sets such as knowledge bases, technical documentation, or datasets. The Files API will integrate with the code execution tool, enabling Claude to access and process uploaded files directly during code execution and produce files such as charts and graphs as part of the response. This means developers can upload a dataset through the Files API once, then have Claude analyze it across multiple sessions without re-uploading. Extended prompt caching Developers can now choose between our standard 5-minute time to live (TTL) for prompt caching or opt for an extended 1-hour TTL at an additional cost —a 12x improvement that can reduce expenses for long-running agent workflows. With extended caching, customers can provide Claude with extensive background knowledge and examples while reducing costs by up to 90% and latency by up to 85% for long prompts. This makes it practical to build agents that maintain context over extended periods, whether they're handling multi-step workflows, analyzing complex documents, or coordinating with other systems. Long-running agent applications that previously faced prohibitive costs can now operate efficiently at scale. Getting started All of these features are now available in public beta on the Anthropic API. Visit our documentation to learn more or watch the keynote from our developer conference to see these capabilities in action.
Explore more product news and best practices for teams building with Claude.
Transform how your organization operates with Claude
Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.
Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.
采集自 Claude Blog，由 collect_claude_blog.py 自动采集
Claude Sonnet 4 now supports 1M tokens of context¶
来源: Claude Blog 发布日期: Aug 12, 2025 采集时间: 2026-05-17 价值评分: 9.0/10 正文字数: ~2021 字符
摘要¶
Claude Sonnet 4 now supports up to 1 million tokens of context—a 5x increase that lets you process entire codebases, synthesize extensive document sets, and build agents that maintain coherence across hundreds of tool calls.
正文内容¶
Claude Sonnet 4 now supports up to 1 million tokens of context—a 5x increase that lets you process entire codebases, synthesize extensive document sets, and build agents that maintain coherence across hundreds of tool calls.
Update: Now available on Google Cloud's Vertex AI (Aug 26, 2025) Claude Sonnet 4 now supports up to 1 million tokens of context on the Anthropic API—a 5x increase that lets you process entire codebases with over 75,000 lines of code or dozens of research papers in a single request. Long context support for Sonnet 4 is now in public beta on the Claude Developer Platform natively, and in Amazon Bedrock and Google Cloud’s Vertex AI. Longer context, more use cases With longer context, developers can run more comprehensive and data-intensive use cases with Claude, including: Large-scale code analysis: Load entire codebases including source files, tests, and documentation. Claude can understand project architecture, identify cross-file dependencies, and suggest improvements that account for the complete system design. Document synthesis: Process extensive document sets like legal contracts, research papers, or technical specifications. Analyze relationships across hundreds of documents while maintaining full context. Context-aware agents: Build agents that maintain context across hundreds of tool calls and multi-step workflows. Include complete API documentation, tool definitions, and interaction histories without losing coherence. API pricing To account for increased computational requirements, pricing adjusts for prompts over 200K tokens: Input Output Prompts ≤ 200K $3 / MTok $15 / MTok Prompts > 200K $6 / MTok $22.50 / MTok
Explore more product news and best practices for teams building with Claude.
Transform how your organization operates with Claude
Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.
Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.
采集自 Claude Blog，由 collect_claude_blog.py 自动采集
Agent view in Claude Code¶
来源: Claude Blog 发布日期: May 11, 2026 采集时间: 2026-05-17 价值评分: 9.0/10 正文字数: ~890 字符
摘要¶
Today we're introducing agent view in Claude Code: one place to manage all your Claude Code sessions.
正文内容¶
Today we're introducing agent view in Claude Code: one place to manage all your Claude Code sessions. When running agents in parallel before, you've probably had to manage multiple terminal tabs, a tmux grid, and an overloaded mental ledger of what you need to tackle next. With agent view in Claude Code, you can kick off new agents, send them to the background, and jump in only when Claude needs you. See at a glance which agents are waiting on you, which are still working, and which are done, so you can easily steer many all at once.
Explore more product news and best practices for teams building with Claude.
Transform how your organization operates with Claude
Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.
Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.
采集自 Claude Blog，由 collect_claude_blog.py 自动采集
Claude on Amazon Bedrock now available to every AWS customer¶
来源: Claude Blog 发布日期: Sep 28, 2023 采集时间: 2026-05-17 价值评分: 9.0/10 正文字数: ~4397 字符
摘要¶
Claude is now generally available on Amazon Bedrock. Claude is now generally available on Amazon Bedrock , the fully managed service that provides Amazon Web Services (AWS) customers with secure cloud access to foundation models and tools for building generative AI applications.
正文内容¶
Claude is now generally available on Amazon Bedrock.
Claude is now generally available on Amazon Bedrock , the fully managed service that provides Amazon Web Services (AWS) customers with secure cloud access to foundation models and tools for building generative AI applications. This means that every AWS customer can now build with Claude, and will soon gain access to an exciting roadmap of new experiences - including Agents for Amazon Bedrock , which our team has been instrumental in developing. Currently available in preview, Agents for Amazon Bedrock can orchestrate and perform API calls using the popular AWS Lambda functions. Through this feature, Claude can take on a more expanded role as an agent to understand user requests, break down complex tasks into multiple steps, carry on conversations to collect additional details, look up information, and take actions to fulfill requests. For example, an e-commerce app that offers a chat assistant built with Claude can go beyond just querying product inventory – it can actually help customers update their orders, make exchanges, and look up relevant user manuals. We also recently shared that we’ll offer secure customization and fine-tuning of Claude models through the service. This technique optimizes Claude’s performance with AWS customers’ expert knowledge and proprietary data to drive more relevant results, while limiting the potential for harmful outputs. The general availability of Claude on Amazon Bedrock advances our work in helping enterprises responsibly integrate transformative AI. In the coming months, we’ll be announcing even more application layer solutions to help organizations get the most value out of Claude. Since announcing our integration with Amazon Bedrock in April, Claude has seen strong organic adoption from leading enterprises including LexisNexis , Lonely Planet and Ricoh USA . Premier asset management firm Bridgewater Associates is also leveraging our state-of-the-art model Claude to create a novel, AI-enabled Investment Analyst Assistant to support junior members of its team. Bridgewater seeks to understand the economy in a fundamental way, uncovering the timeless and universal cause-and-effect linkages that drive global markets. In order to reveal those relationships, analysts at Bridgewater spend significant time understanding and systematically stress testing hypotheses, often in code. For example, an analyst might investigate the impact of the Federal Reserve’s tightening of commodity prices by creating key data visualizations (including charts and tables) to test hypotheses, synthesize findings, and share their research. Bridgewater's team is working with Claude on Amazon Bedrock to make that process more efficient and scalable. “Claude is the highly capable model behind our upcoming Investment Analyst Assistant on Amazon Bedrock, which is able to take basic instructions, generate Python code, work through errors, and output charts and tables much like a first or second-year analyst would”, said Aaron Linsky, CTO - AI/ML at Bridgewater Associates. “When the Investment Analyst Assistant isn’t certain about the path to take, it will also ask clarifying questions before attempting to execute.” Linsky continued, “Our Investment Analyst Assistant also enables the editing of code in place, which lets our analysts choose when to leverage Claude, or when to make the simple change themselves. This enables quicker loops and lets the Investment Analyst Assistant iterate on the updated codebase; something that many other chat-style assistants simply don’t offer.” Getting started with Claude on Amazon Bedrock We’re pleased to provide even more of the AWS community with access to Claude, and to contribute our technical team’s expertise towards responsible innovation on the Amazon Bedrock service. Get started with Claude on Amazon Bedrock today. Already set up on Amazon Bedrock? Visit our documentation page for developer resources, prompt engineering guidance, and more. We can’t wait to see what you build.
Explore more product news and best practices for teams building with Claude.
Transform how your organization operates with Claude
Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.
Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.
采集自 Claude Blog，由 collect_claude_blog.py 自动采集
Introducing the analysis tool in Claude.ai¶
来源: Claude Blog 发布日期: Oct 24, 2024 采集时间: 2026-05-17 价值评分: 9.0/10 正文字数: ~1998 字符
摘要¶
Claude can now write and run JavaScript code directly in Claude.ai through the analysis tool, enabling precise data analysis and real-time insights with mathematically accurate results.
正文内容¶
Claude can now write and run JavaScript code directly in Claude.ai through the analysis tool, enabling precise data analysis and real-time insights with mathematically accurate results.
Update: The analysis tool is being replaced by more powerful code execution capabilities. (November 5, 2025) Code execution can run all the same analysis, plus generate downloadable files (spreadsheets, CSVs, reports), create visualizations, and handle complex multi-step workflows. Learn more. We're introducing the analysis tool, a new built-in feature for Claude.ai that enables Claude to write and run JavaScript code. Claude can now process data, conduct analysis, and produce real-time insights. The analysis tool is available for all Claude.ai users in feature preview . Think of the analysis tool as a built-in code sandbox, where Claude can do complex math, analyze data, and iterate on different ideas before sharing an answer. The ability to process information and run code means you get more accurate answers—building on Claude 3.5 Sonnet’s state-of-the-art coding and data skills. Analyzing and visualizing data from CSV files When you need precise, verifiable answers from data, Claude now works more like a real data analyst. Instead of relying on abstract analysis alone, it can systematically process your data—cleaning, exploring, and analyzing it step-by-step until it reaches the correct result. While Claude could always write code, it can now run that code within Claude.ai to support all types of analysis tasks. With the analysis tool, you get answers that are not just well-reasoned, but are mathematically precise and reproducible.
Explore more product news and best practices for teams building with Claude.
Transform how your organization operates with Claude
Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.
Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.
采集自 Claude Blog，由 collect_claude_blog.py 自动采集