Skip to content

About

Free read-only AI/LLM API cost review for GitHub: find token, retry, cache, and model-call signals without uploading source.

Topics

Resources

Contributing

Security policy

Stars

2 stars

Watchers

0 watching

Forks

Use this GitHub action with your project
Add this Action to an existing workflow or create a new one
View on Marketplace

Latest commit

 

History

67 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CostDoctor

Free AI/LLM API Cost Review for GitHub — read-only. Enter one public repository URL and receive a Static Precheck plus an evidence-bound deep diagnosis in the same Issue.

Free, read-only AI/LLM cost review for GitHub, with English as the default and Korean available.

English · 한국어 안내

FREE · READ-ONLY · NO API KEY · NO CODE EXECUTION · NO CODE MODIFICATION · EXTERNAL TELEMETRY OFF

Start in seconds

What happens

  1. Enter one public GitHub repository URL.
  2. Choose English (the default) or 한국어.
  3. Submit the trusted GitHub Issue Form.
  4. Receive one CostDoctor comment containing Stage 1 Static Precheck and Stage 2 Deep Diagnosis.

The free path reads a bounded checkout in a temporary GitHub-hosted runner. It does not execute the target project, call an LLM/provider API, request a secret, write to the target repository, or send customer source to an operator computer.

What the result means

The report orders findings as: candidate signal → runtime-impact evidence → safe repository location or call group → measurable token/context effect when available → concrete fix or measurement action → expected effect or why it cannot be estimated → confidence and unknowns.

Static evidence is not a bill, API usage receipt, or verified savings claim. Without matching provider usage, price, quality, and Before/After evidence, cost and savings stay UNKNOWN. Test, evaluation, documentation, inactive-retry, and ordinary-file-cache signals are not promoted to production cost loss without runtime support.

Private repositories

Never paste a private URL into the public form. Use Private Self-Scan inside your own GitHub Actions workflow.

Boundary Behavior
Permission contents: read by default
Location Your GitHub-hosted runner's temporary workspace
Writes No commit, push, branch, PR, or merge
Provider calls No API key or paid model required for structural diagnosis
Telemetry External telemetry is off by default

Repeat checks with the Action

permissions:
  contents: read

steps:
  - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
    with:
      persist-credentials: false
  - name: Review AI cost signals
    uses: leesugwan-dot/cost-doctor-github-app@v1.0.1

The Action writes a runner-local report.json and report.md and exposes report-directory and scan-status outputs. See the Action start guide.

Evidence and safety

  • No automatic code fixes, commits, branches, pull requests, merges, pricing, or billing features are enabled.
  • Public Scan and Private Self-Scan are read-only and fail closed on unsafe or incomplete input.
  • Original source, secrets, prompts, private paths, and raw credentials are not placed in public results.
  • Optional verified measurement uses the customer's own provider account, GitHub Secret, and an approved per-workload spend cap.

Evidence-based guides

Learn more

Public Scan · Private Self-Scan · Sanitized result example · Quick Start · FAQ · Rollback · Feedback

Privacy · Terms · Operator policy · Security · Support · Apache-2.0 license

CostDoctor is Apache-2.0 at the public integration layer. Automatic code changes, paid features, and external-provider automatic fixes are not advertised as available features.

About

Free read-only AI/LLM API cost review for GitHub: find token, retry, cache, and model-call signals without uploading source.

Topics

Resources

Contributing

Security policy

Stars

2 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages