Claude Agent SDK Gets a Monthly Credit on Paid Claude Plans Starting June 15, 2026

Anthropic published a support article confirming that paid Claude plans get a dedicated monthly credit for programmatic Claude usage starting June 15, 2026. The credit is plan-tier-specific, opt-in, and covers Claude Agent SDK projects plus a defined set of Claude Code surfaces. (Anthropic support)

What changes on June 15, 2026

Anthropic's support article gives the effective date as June 15, 2026, and limits eligibility to Pro, Max, Team, and Enterprise subscription plans. API-key-only users are explicitly excluded -- the credit is tied to a Claude subscription. (Anthropic support)

Until now, building on the Agent SDK or running claude -p in CI meant either burning your plan's interactive Claude Code allowance or paying API rates on a separate API key. The June 15 change carves out a programmatic budget inside the same subscription, so plan holders no longer have to choose between agent automation and the IDE work that pays for the subscription in the first place.

Monthly credit by plan

The support article lists the per-plan credit allowance:

Source: Anthropic support article on using the Claude Agent SDK with your Claude plan.

Credits are per-user and non-poolable -- a Team Premium seat gets its own $100 per month, and that allowance does not roll up into a shared pool. Unused credits do not roll over between billing cycles. (Anthropic support)

What counts as programmatic usage

The support article scopes the credit to four programmatic surfaces:

• Claude Agent SDK usage in your own Python or TypeScript projects.
• The claude -p command in Claude Code (non-interactive mode).
• Claude Code GitHub Actions integration.
• Third-party apps authenticating through the Agent SDK.

Interactive Claude Code sessions and Claude.ai conversations are not part of this credit -- they continue to draw from your existing plan allowance. (Anthropic support)

Activating the credit

Anthropic describes activation as a one-time opt-in: claim the credit through your account once, and "after that, it refreshes automatically each cycle." (Anthropic support)

The user-visible consequence is that subscribers who never opt in get no Agent SDK credit even though the plan technically qualifies. If you are running Agent SDK code on June 15 or later, the first action is to confirm the credit is claimed for your account.

What happens when the credit runs out

Anthropic spells out two paths after the monthly credit is exhausted:

• If usage credits are enabled on the account, "additional Agent SDK usage flows to usage credits at standard API rates."
• If usage credits are not enabled, "Agent SDK requests stop until your credit refreshes." (Anthropic support)

That distinction matters operationally. A CI job built on Claude Code GitHub Actions will fail mid-pipeline once the monthly credit runs out unless the account has usage credits enabled as a fallback. Teams running scheduled agents or PR-bots should confirm the usage-credit toggle before June 15.

Why it matters

The Agent SDK credit normalizes a billing pattern that the market has been moving toward all year: a programmatic budget bundled into a developer subscription. It removes the "do I run this on my plan or pay API rates" decision for low-to-mid-volume agent work, and it pushes the breakeven point at which moving to raw API billing makes sense further out.

For agencies, contractors, and indie builders who already pay for Max, the practical effect is that an extra $100-$200 of Agent SDK usage now ships inside the seat they were already paying for. For Team and Enterprise buyers, per-seat credits give a predictable lower bound on programmatic spend without enabling unlimited usage credits.

Sources

• Anthropic support: Use the Claude Agent SDK with your Claude plan -- effective date, plan eligibility, credit table, activation, and overage behavior.
• @ClaudeDevs on X -- announcement post linking to the support article.