Paid Claude subscriptions now include monthly Agent SDK credits for programmatic agents

Anthropic has changed how some paid Claude subscriptions handle developer-facing agent usage. A Help Center article updated on May 16 says users on Pro, Max, Team and some Enterprise plans can now claim a separate monthly credit for Claude Agent SDK usage, the `claude -p` command, GitHub Actions integration, and certain third-party apps built on the Agent SDK. The same article says the change takes effect June 15, 2026, and that this usage will no longer count against a subscription’s normal Claude plan limits. The update shifts Agent SDK usage into a separate bucket rather than bundling it with interactive Claude use. Anthropic says subscription usage limits for Claude on the web, desktop and mobile apps, along with interactive Claude Code and Claude Cowork, stay the same and remain reserved for those interactive products. (support.claude.com) ### Which Claude plans now get the monthly credit? Anthropic’s Help Center says the credit is available on Pro, Max, Team and Enterprise plans starting June 15, 2026. The article lists monthly credit amounts of $20 for Pro, $100 for Max 5x, $200 for Max 20x, $20 for Team standard seats, $100 for Team premium seats, $20 for usage-based Enterprise, and $200 for seat-based Enterprise premium seats. (support.claude.com) It also says seat-based Enterprise standard seats are not eligible. The same page says the credit is per-user rather than pooled across a team. Anthropic says unused credit does not roll over and resets with each billing cycle after a one-time opt-in through the user’s Claude account. ### What exactly does the credit pay for? (support.claude.com) Anthropic says the monthly credit applies to four categories: Claude Agent SDK usage in a user’s own Python or TypeScript projects, the non-interactive `claude -p` command in Claude Code, the Claude Code GitHub Actions integration, and third-party apps that authenticate with a Claude subscription through the Agent SDK. (support.claude.com) The Help Center article also draws a boundary around what is not included. Interactive Claude Code sessions in the terminal or IDE, Claude conversations in consumer apps, Claude Cowork, and other features that use extra usage are excluded and continue to draw from the existing subscription limits or other billing paths, Anthropic says. (support.claude.com) ### What happens after the credit runs out? Anthropic says Agent SDK usage draws from the monthly credit first. After that, the company says usage moves to extra usage at standard API rates only if the customer has enabled extra usage; otherwise Agent SDK requests stop until the next billing-cycle refresh. (support.claude.com) A separate Help Center article on Claude Code already warned that if a user has an `ANTHROPIC_API_KEY` environment variable set, Claude Code will authenticate with that API key instead of the Claude subscription, which can result in API charges rather than subscription-based usage. (support.claude.com) ### How does this fit with Anthropic’s developer stack? Anthropic’s Agent SDK documentation describes the SDK as a way to build AI agents in Python and TypeScript using the same tools, agent loop and context management that power Claude Code. The docs say the SDK includes built-in tools for reading files, running commands and editing code. (support.claude.com) Anthropic’s broader developer docs describe two routes for builders: the Messages API for direct prompting access and Claude Managed Agents as a managed harness for long-running tasks and asynchronous work. That places the new subscription credit alongside Anthropic’s wider push to make agent-style workflows available both through code libraries and managed infrastructure. (docs.claude.com) ### Does this mean developers can log into third-party apps with Claude subscriptions? Anthropic’s documentation shows a tension between the new credit policy and earlier SDK guidance. The Help Center now says the monthly credit can cover “third-party apps that authenticate with your Claude subscription through the Agent SDK.” (docs.claude.com) But the current Agent SDK overview also says that, unless previously approved, Anthropic does not allow third-party developers to offer `claude.ai` login or Claude rate limits for their products, including agents built on the SDK, and tells developers to use API-key authentication methods instead. Anthropic has not, in the documents reviewed, explained how broadly the Help Center language changes that earlier restriction or whether approvals are still required. (support.claude.com) ### Where can users check the details and next steps? Anthropic’s Help Center article says eligible users must claim the credit once through their Claude account, after which it refreshes automatically every billing cycle. The same page says Claude Platform accounts using an API key do not receive the credit and that pay-as-you-go billing continues as before. (docs.claude.com) Anthropic’s Agent SDK docs continue to point developers to the Python and TypeScript packages and quickstart materials, while the company’s broader docs position Managed Agents for asynchronous and long-running workloads. Those pages, together with the updated Help Center article dated May 16 and effective June 15, 2026, are the main references for developers tracking the rollout. (support.claude.com)