Clarify quota/billing behavior for Claude Agent SDK runtime
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 48/100
- Issue type
- Documentation
- Clarity
- Mostly clear
- Activity status
- Quiet
- Tech stack
- typescript
- Domain
- developer-experience, documentation
Research direction
Start with the README, package.json, src/lib/claude-client.ts, and src/lib/runtime/sdk-runtime.ts to trace which runtime and authentication paths the issue describes. Verify the quota behavior against the linked Anthropic documentation, then clarify the supported billing sources in the README and runtime selection UI; done means users can tell which account or quota their chosen runtime uses.
Written by the indexing model from the issue text.
Description
Hi, thanks for building CodePilot. I am evaluating the runtime and remote bridge behavior, and noticed one billing/quota detail that may be worth documenting more explicitly.
CodePilot appears to support a Claude Code / Claude Agent SDK runtime via @anthropic-ai/claude-agent-sdk, while the README also says users can configure providers and track token usage/costs. Anthropic's current docs say that, starting June 15, 2026, Claude Agent SDK and claude -p usage on Claude subscription plans will draw from a separate monthly Agent SDK credit. API key users continue with normal pay-as-you-go API billing.
Relevant docs:
- https://code.claude.com/docs/en/agent-sdk/overview
- https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan
CodePilot references I checked:
package.jsondepends on@anthropic-ai/claude-agent-sdksrc/lib/claude-client.tsimportsqueryfrom@anthropic-ai/claude-agent-sdksrc/lib/runtime/sdk-runtime.tsexposes theclaude-code-sdkruntime- README says Claude Code CLI is optional for basic chat and mentions token/cost tracking
Questions:
- When using the Claude Agent SDK runtime with Claude subscription auth, does usage count against the separate Agent SDK monthly credit?
- When CodePilot injects a configured provider API key/base URL into the SDK subprocess, does usage count against that provider/API billing instead?
- Are remote Bridge sessions, such as Feishu/Telegram/Discord turns, counted the same way as local chat when they use the SDK runtime?
- Could the UI or README show the current runtime's quota source, for example:
- Native provider runtime -> configured provider/API key billing
- Claude Agent SDK via Claude subscription -> Agent SDK monthly credit
- Claude Agent SDK via API key/provider -> API/provider billing
- Local interactive Claude Code/channel-style runtime, if supported -> different from Agent SDK /
claude -p
Product-level concern:
A user may try CodePilot expecting their existing Claude Max subscription to provide a broad remote Claude Code experience across Feishu/Telegram/Discord/etc. But if the advanced or remote workflow depends mainly on the Claude Agent SDK runtime, that experience may be constrained by a separate Agent SDK credit and may require additional usage credits or API billing after that credit is exhausted.
That could materially affect user expectations, adoption, and retention. Users might assume all CodePilot usage is billed through the configured provider or through normal Claude Code interactive limits, while Agent SDK usage may draw from a different quota/billing source.
It may be helpful to clarify this in the README and runtime selection UI, and possibly add a small warning/label showing which account/quota/billing source the current runtime will consume.
Apologies if I misunderstood part of the runtime routing. I am filing this mainly as a docs/UX clarification because the quota source seems important for users evaluating the product.
- Dominant language
- TypeScript
- Stars
- 6.5k
- Forks
- 697
- PR merge metrics
- No merged PRs in 30d
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- No contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from op7418/CodePilot
-
发生错误时提醒Error:successOpenbug P0-crash-or-interrupt
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 75/100
-
Difficulty 2/5 1-3 hours Newbie friendliness 65/100
-
Effort selector dropdown hidden for GLM models — blocks access to GLM-5.2 High/Max effort tiersOpen
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
-
bug P0-crash-or-interrupt
Difficulty 4/5 3-5 days Newbie friendliness 48/100
All issues in op7418/CodePilot
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 70/100
backnotprop/plannotator#2914 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
microsoft/vscode-containers#655 ·
Maintainers usually reply within 1 day
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 77/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
dsh-tauri-desk/deepseek-harness-desktop#1477 · 1 comment ·