🐞 GitHub Issue Insights

🚀 Overview

GitHub Issue Insights is an Apify Actor that transforms GitHub issues into clear, structured, and multilingual insights.

🎯 Who Is This For?

• Contributors looking for actionable issues (including Good First Issues) in their native language.
• Maintainers who want quick clarity on decisions, gaps, and next steps
• Global Teams bridging language barriers in open-source collaboration.
• Project Managers needing quick summaries of complex issues.

---

✨ Features

• 📥 Reliable Issue Fetching: Uses GitHub API. To retrieve issue title, description, comments, and linked Pull Requests from any public GitHub repository.

• 🧠 AI Analysis: Leveraging Google Gemini, it extracts:

  • Issue category (Bug / Feature / Improvement)
  • Severity level
  • Technical context (language, framework) with confidence scores
  • Contributor difficulty & required skills
  • Decisions made, unresolved gaps, and open questions
  • Clear, actionable next steps

• 🌍 Multilingual Insights: Using Lingo.dev

  • Human-readable fields are translated using Lingo.dev
  • Technical metadata, keys, URLs, and labels remain unchanged
  • Supports languages like Spanish, French, Hindi, and more

• ✅ Structured Output:

  • Returns a well-defined JSON schema
  • Designed for automation, dashboards, and further processing -Safe defaults with explicit inference disclaimers

---

🛠️ Input Parameters

Field	Type	Description	Required
issue_url	String	Full URL of a public GitHub issue	✅
target_language	String	ISO language code (en, es, fr, hi, etc.)	✅
gemini_api_key	String	API key from Google AI Studio	✅
lingo_api_key	String	API key from Lingo.dev	✅

---

📦 Output Example

Below is a shortened example of the structured output (translated to Spanish):

{
  "issue_data": {
    "state": "open",
    "assignees": [],
    "linked_pull_requests": []
  },
  "issue_analysis": {
    "type": "Bug",
    "severity": "High",
    "labels": ["bug", "urgent"]
  },
  "tech_context": {
    "language": "JavaScript",
    "framework": "React",
    "confidence": 0.95
  },
  "summary": {
    "issue": "El usuario reporta un fallo al hacer clic en el botón de guardar...",
    "discussion": "Varios desarrolladores han confirmado el error en la versión 2.1..."
  },
  "contributor_info": {
    "good_first_issue": true,
    "difficulty": "Low",
    "skills": ["React", "CSS"]
  },
  "decisions_and_gaps": {
    "decisions": ["Se decidió revertir el commit anterior."],
    "gaps": ["Faltan pruebas unitarias para este componente."],
    "open_questions": ["¿Afecta esto a la versión móvil?"]
  },
  "next_steps": [
    "Reproducir el error en el entorno local.",
    "Crear un caso de prueba fallido."
  ],
  "language": "es"
}

---

⚙️ How It Works

1. Fetch: Parses the issue URL and retrieves issue data and comments using the GitHub API.
2. Analyze: Sends the issue content to Google Gemini, prompting it to extract structured insights as a senior technical reviewer.
3. Translate: Translates only human-readable fields using Lingo.dev, preserving technical accuracy.
4. Deliver: Stores the final structured output in the Apify dataset.

---

📝 Notes

• Only public repositories are supported
• Private repositories require authentication (not supported in this version)
• Analysis quality depends on the clarity of the issue and discussion
• Repository files are not read
• Technical context is inferred and includes confidence scores
• Linked pull requests are provided as references only
• Each output explicitly includes:
  • data_sources
  • inference_notice