Dynamic Workflows in Claude Code
Recorded: May 28, 2026, 6 p.m.
| Original | Summarized |
Introducing dynamic workflows | Claude Meet ClaudeProductsClaudeClaude CodeClaude CoworkClaude SecurityFeaturesClaude for ChromeClaude for SlackClaude for Microsoft 365SkillsModelsOpusSonnetHaikuPlatformOverviewDeveloper docsPricingConsole loginSolutionsUse casesAI agentsCodingCompany sizeStartupsEnterpriseDepartmentsLegalSecurityIndustriesCustomer supportEducationFinancial servicesGovernmentHealthcareLife sciencesNonprofitsPricingOverviewAPIPlansProMaxTeamEnterpriseResourcesInsightsBlogCustomer storiesAnthropic newsLearnAnthropic AcademyCoursesTutorialsUse casesToolsConnectorsPluginsConnectEventsCommunityLoginContact salesContact salesContact salesTry ClaudeTry ClaudeTry ClaudeContact salesContact salesContact salesTry ClaudeTry ClaudeTry ClaudeContact salesContact salesContact salesTry ClaudeTry ClaudeTry ClaudeContact salesContact salesContact salesTry ClaudeTry ClaudeTry ClaudeMeet ClaudeProductsClaudeClaude CodeClaude CoworkClaude SecurityFeaturesClaude for ChromeClaude for SlackClaude for Microsoft 365SkillsModelsOpusSonnetHaikuPlatformOverviewDeveloper docsPricingConsole loginSolutionsUse casesAI agentsCodingCompany sizeStartupsEnterpriseDepartmentsLegalSecurityIndustriesCustomer supportEducationFinancial servicesGovernmentHealthcareLife sciencesNonprofitsPricingOverviewAPIPlansProMaxTeamEnterpriseResourcesInsightsBlogCustomer storiesAnthropic newsLearnAnthropic AcademyCoursesTutorialsUse casesToolsConnectorsPluginsConnectEventsCommunityLoginContact salesContact salesContact salesTry ClaudeTry ClaudeTry ClaudeContact salesContact salesContact salesTry ClaudeTry ClaudeTry ClaudeBlogBlog/Introducing dynamic workflows in Claude CodeExplore hereAsk questions about this pageCopy as markdownIntroducing dynamic workflows in Claude CodeCategoryProduct announcementsProductClaude CodeDateMay 28, 2026Reading time5minShareCopy linkhttps://claude.com/blog/introducing-dynamic-workflows-in-claude-codeToday we're introducing dynamic workflows in Claude Code, helping Claude take on the most challenging tasks end-to-end. Work you'd normally plan in quarters now finishes in days. Claude dynamically writes orchestration scripts that run tens to hundreds of parallel subagents in a single session, checking its work before anything reaches you.Some problems are too big for one pass by a single agent, especially in complex, legacy codebases: a bug hunt across an entire service, a migration that touches hundreds of files, a plan you want stress-tested from every angle before you commit to it. Dynamic workflows can handle all of these end-to-end.Dynamic workflows are available today in research preview in the Claude Code CLI, Desktop, and the VS code extension for Max, Team, and Enterprise (if admin enabled) plans, as well as on the Claude API, on Amazon Bedrock, Vertex AI, and Microsoft Foundry.Note: Dynamic workflows can consume substantially more tokens than a typical Claude Code session, so we recommend starting on a scoped task to get a feel for usage in your work. For the best experience, turn on auto mode when using dynamic workflows. From there, you have two ways to start a workflow:Ask Claude to create a dynamic workflow directly (e.g., “Create a workflow”), orSwitch on a new Claude Code-specific setting called ultracode. This is accessible through the effort menu and it sets the effort level to xhigh, while letting Claude decide automatically when to use a workflow to handle your task.Dynamic workflows in actionEarly access users and teams inside Anthropic have been using dynamic workflows for a wide range of use cases, including:Codebase-wide bug hunts, profiler-guided optimization audits, and security audits: Claude searches a service or repo in parallel, then runs independent verification on every finding so the report surfaces real issues. The same shape works for hardening passes: auth checks, input validation, and unsafe patterns across an entire codebase.Large migrations and modernization efforts: Claude can handle framework swaps, API deprecations, language ports that span thousands of files end-to-end.Critical work you need checked twice: When the cost of a wrong answer is high, a workflow gives Claude independent attempts at the problem and adversarial agents working to break the result before you see it.“Dynamic workflows have been especially valuable for discovery and review tasks across large codebases. We’ve seen strong results using it to identify dead code and surface cleanup opportunities that traditional static analysis missed, helping our engineers move faster on maintenance and refactoring work.”Alessio Vallero, Senior Engineering Manager“Dynamic workflows fill the gap between firing off a single subagent and building out a full agent team. Plan to implementation just flows, so we can trust longer runs without losing visibility.”Ken Takao, Lead Systems EngineerPrevPrev0/5NextNexteBookRewriting Bun with dynamic workflowsAn example of what dynamic workflows can unlock at scale is the recent rewrite of Bun. Jarred Sumner used dynamic workflows to port Bun from Zig to Rust with 99.8% of the existing test suite passing, roughly 750,000 lines of Rust, and eleven days from first commit to merge. One workflow mapped the right Rust lifetime for every struct field in the Zig codebase. The next wrote every .rs file as a behavior-identical port of its .zig counterpart, hundreds of agents working in parallel with two reviewers on each file. A fix loop then drove the build and test suite until both ran clean. After the port landed, an overnight workflow addressed unnecessary data copies and opened a PR for each for final review. While not yet in production, all of this was handled by dynamic workflows. Jarred will be writing about this more in the future.How it worksWhen a workflow kicks off, Claude plans dynamically based on your prompt, breaks it into subtasks, and fans the work out across subagents running in parallel. Results are checked before they're folded in, and you come back to a single, coordinated answer. Agents address the problem from independent angles, other agents try to refute what they found, and the run keeps iterating until the answers converge—which is how a workflow reaches results a single pass can't.Dynamic workflows are built for parallel and long-running work that can extend into hours and days, doing the most complex engineering work that previously would have taken weeks. Progress is saved as the run goes, so a job that's interrupted picks up where it left off instead of starting over. Because the coordination happens outside the conversation, the plan stays on track no matter how big the task gets.It’s important to note that dynamic workflows consume meaningfully more usage than a typical Claude Code session. The first time a workflow triggers, Claude Code shows what's about to run and asks you to confirm. Organization admins can also optionally disable workflows through managed settings.Getting startedIf you're on a Max or Team plan, or using Claude Code via the API, dynamic workflows are on by default. Ask Claude to create a workflow or turn on the Claude Code-specific setting ultracode to get started. If you’re on an Enterprise plan, dynamic workflows are off by default at launch. Your admin can easily change this in the Claude Code settings.Read the documentation to learn more.FAQNo items found.Get Claude CodeDesktopVS CodeJetBrainsOn the webSlack Cookie settings We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here. Customize cookie settings Reject all cookies Accept all cookies Necessary Required Analytics Off Marketing Off Save preferences Privacy policyPrivacy policyPrivacy policyResponsible disclosure policyResponsible disclosure policyResponsible disclosure policyTerms of service: CommercialTerms of service: CommercialTerms of service: CommercialTerms of service: ConsumerTerms of service: ConsumerTerms of service: ConsumerUsage policyUsage policyUsage policyx.comx.comLinkedInLinkedInYouTubeYouTubeInstagramInstagramEnglish (US)English (US)日本語 (Japan)Deutsch (Germany)Français (France)한국어 (South Korea) |
Dynamic workflows in Claude Code are introduced to enable Claude to manage and execute complex tasks end-to-end, effectively compressing work that typically spans quarters into a matter of days. These workflows allow Claude to dynamically generate orchestration scripts that run tens to hundreds of parallel subagents within a single session, incorporating a verification step to ensure accuracy before presenting results to the user. This capability addresses challenges presented by large, intricate tasks, such as locating bugs across entire service codebases, executing migrations involving numerous files, or stress-testing plans from multiple perspectives. Dynamic workflows are accessible today in research preview through the Claude Code CLI, Desktop applications, and the VS code extension for users on the Max, Team, and Enterprise plans, as well as via the Claude API, Amazon Bedrock, Vertex AI, and Microsoft Foundry. Users are cautioned that dynamic workflows may consume a significantly larger amount of tokens than a standard Claude Code session; therefore, starting with a scoped task is recommended to gauge usage. For optimal experience, utilizing auto mode or switching on the Claude Code specific setting, ultracode, allows Claude to automatically determine when to deploy a workflow for a given task. These workflows have demonstrated utility across various advanced engineering scenarios. They can facilitate tasks like codebase-wide bug hunts, profiler-guided optimization audits, and security audits by having Claude perform parallel searches and independent verification across code and findings. Furthermore, they are valuable for handling large-scale migration and modernization efforts, including framework swaps, API deprecations, and language porting that affect thousands of files. Dynamic workflows also address critical work that requires rigorous scrutiny by allowing Claude to execute independent attempts and utilize adversarial agents to challenge and refine the solution before a final result is delivered. The value of these workflows is particularly evident in discovery and review tasks across extensive codebases, where they have helped identify dead code and surface cleanup opportunities that traditional static analysis methods might miss, thereby accelerating maintenance and refactoring work, as noted by Alessio Vallero, Senior Engineering Manager. Dynamic workflows bridge the gap between deploying a single subagent and constructing a complete agent team, providing the necessary visibility for longer, more complex operations. An illustrative example of this capacity is the recent effort to rewrite the Bun project, where Jarred Sumner utilized dynamic workflows to port the project from Zig to Rust, successfully managing the migration of approximately 750,000 lines of Rust code while ensuring 99.8 percent of the existing test suite passed. This involved multiple agents working in parallel, performing sophisticated mappings and subsequent fix loops that iterated until the build and test suites were clean, demonstrating a capacity for handling parallel, iterative engineering tasks that previously required weeks. The operational mechanism of a dynamic workflow involves Claude dynamically planning the task, deconstructing it into subtasks, and distributing the work across parallel subagents. The system incorporates intermediate checks on results, ensuring coordinated output. Agents tackle problems from disparate angles, engage in antagonistic testing, and iterate until the solutions converge, thereby achieving results that a single sequential pass cannot attain. These workflows are specifically designed for parallel and long-running operations, capable of extending over hours or days, with progress automatically saved to allow for resumption if the job is interrupted, ensuring the planned trajectory remains intact regardless of task complexity. |