{"slug": "open-source-toolkit-to-analyze-your-chatgpt-claude-usage-from-exports", "title": "Open source toolkit to analyze your ChatGPT/Claude usage from exports", "summary": "A new open-source toolkit enables users to analyze their personal ChatGPT and Claude usage data entirely offline, processing official export files to generate model adoption timelines, topic breakdowns, and prompt engineering effectiveness metrics. The tool, available on GitHub, also computes cost efficiency metrics like Productive Output per Dollar (POD) and includes three methodology papers for quantifying practice hours and usage profiles. All analysis runs locally with no data leaving the user's machine, addressing privacy concerns around AI usage tracking.", "body_md": "Privacy-first tools for analyzing your personal LLM usage from official ChatGPT and Claude exports. Everything runs locally. Nothing leaves your machine.\n\nTurn your raw export data into: model adoption timelines, topic breakdowns, prompt engineering effectiveness metrics, cost efficiency analysis, and formatted reports.\n\n```\ngit clone https://github.com/noah-chelednik/llm-export-analytics.git\ncd llm-export-analytics\n./run_pipeline.sh --sample\n```\n\nThis runs the full pipeline against included sample data so you can see what it produces before using your own exports.\n\n**Model adoption timeline** shows which AI models you used and when you switched:\n\n```\n2025-01  ████████████████████████████████ GPT-4o (100%)\n2025-02  ██████████████████████████ GPT-5 (66%)  ████ GPT-4o (21%)\n2025-03  ████████████████████████ GPT-5.2-T (78%)  ████ GPT-5 (22%)\n```\n\n**Prompt technique tracking** measures how your prompting style evolves over time:\n\n| Technique | Adoption Rate | First Appeared |\n|---|---|---|\n| Constraint specification | 6.5% | 2023-08 |\n| Context front-loading | 4.6% | 2024-02 |\n| Code inclusion | 2.7% | 2023-10 |\n| Multi-step instructions | 1.7% | 2023-11 |\n\n**Cost efficiency** computes your Productive Output per Dollar (POD) from subscription costs and total output words. Also computes quality-adjusted POE with sensitivity analysis.\n\n**Plus:** topic distribution, session dynamics, interaction style breakdown, conversation outcome classification, industry benchmark comparison, and more.\n\nThree methodology papers are included:\n\n-\n: Rigorous methodology for quantifying LLM practice hours from export data. Tiered claims with sensitivity analysis and stress testing.[LLM_Practice_Hours_Methodology_GIT.pdf](/noah-chelednik/llm-export-analytics/blob/main/LLM_Practice_Hours_Methodology_GIT.pdf) -\n: Comprehensive data-driven usage profile covering model adoption (21 model versions), domain portfolio, prompt engineering effectiveness, and interaction patterns.[Deep_LLM_Usage_Analysis.pdf](/noah-chelednik/llm-export-analytics/blob/main/docs/Deep_LLM_Usage_Analysis.pdf) -\n: The first standardized methodology for measuring individual LLM cost efficiency. Defines Productive Output per Dollar (POD), Deliverable-Linked Output per Dollar (DLOD), and quality-adjusted Productive Output Efficiency (POE).[Productive_Output_Efficiency.pdf](/noah-chelednik/llm-export-analytics/blob/main/docs/Productive_Output_Efficiency.pdf)\n\n```\nscripts/\n  analyze_chatgpt.py              # ChatGPT export normalization and basic stats\n  analyze_claude.py               # Claude export normalization and basic stats\n  analyze_combined.py             # Combined cross-platform analysis\n\nscripts/deep_analysis/\n  extract_chatgpt_metadata.py     # Model versions, tools, branching, reasoning\n  classify_and_link.py            # Topic classification + project attribution\n  analyze_effectiveness.py        # Prompt techniques, outcomes, interaction patterns\n  generate_tables_and_charts.py   # Formatted tables and ASCII charts\n  compute_pod.py                  # Productive Output per Dollar\n  compute_dlod.py                 # Deliverable-Linked Output per Dollar\n  compute_poe.py                  # Quality-adjusted POE with sensitivity analysis\n  compare_benchmarks.py           # Industry benchmark comparison\n\nexamples/\n  sample_chatgpt_export.json      # Synthetic sample data for testing\n  sample_claude_export.json       # Synthetic sample data for testing\n\ntemplates/\n  example_projects.json           # Example project config (customize for your projects)\n  cost_log_template.json          # Template for subscription cost tracking\n  deliverable_inventory_template.json  # Template for verified deliverables\n  quality_params.json             # Q coefficient configurations for POE\n  benchmarks.json                 # Curated industry benchmark data with citations\n\ndocs/\n  DEEP_ANALYSIS_GUIDE.md          # Step-by-step guide to the full pipeline\n  Deep_LLM_Usage_Analysis.pdf\n  Productive_Output_Efficiency.pdf\n  Prompting_One_Pager.pdf\n\nresults/\n  usage_snapshot_2025-12-09.md    # Historical snapshot\n  usage_snapshot_2026-05-26.md    # Current snapshot\n```\n\n**ChatGPT:** Settings > Data controls > Export data. You'll receive a ZIP containing `conversations.json`\n\n(or multiple shards).\n\n**Claude:** Settings > Account > Export data. You'll receive a ZIP containing `conversations.json`\n\n.\n\n```\npython3 -m venv .venv\nsource .venv/bin/activate\npip install -r requirements.txt\npython scripts/analyze_chatgpt.py --input /path/to/conversations.json --out outputs --utc\npython scripts/analyze_claude.py --input /path/to/conversations.json --out outputs --utc\npython scripts/analyze_combined.py \\\n  --chatgpt outputs/chatgpt_messages_normalized.csv \\\n  --claude outputs/claude_messages_normalized.csv --utc\n```\n\nRun the basic scripts with `--include-content`\n\nfirst, then see the [Deep Analysis Guide](/noah-chelednik/llm-export-analytics/blob/main/docs/DEEP_ANALYSIS_GUIDE.md) for the full pipeline: model tracking, topic classification, prompt effectiveness, cost efficiency, and industry benchmarks.\n\nAll processing is local. No network calls. No telemetry.\n\nBy default, scripts produce privacy-minimized CSVs (timestamps, IDs, roles only). Content analysis requires the explicit `--include-content`\n\nflag. Never commit or share CSVs generated with content enabled.\n\n**ChatGPT:** Reconstructs the primary conversation path from the mapping tree. Walks parent pointers from the current node back to root. Extracts per-message model metadata, tool usage, branching, and reasoning information.\n\n**Claude:** Extracts text blocks from content arrays. Normalizes roles to match the shared schema. Supports both web/app and Claude Code CLI sessions.\n\n**Deep analysis:** Adds topic classification via keyword taxonomy with content fallback, optional project attribution from user-provided config, prompt technique detection (10 techniques tracked), conversation outcome classification, interaction style analysis, and cost efficiency computation.\n\n- Results depend on platform export formats which may change\n- ChatGPT analysis reconstructs primary path only, not full conversation trees\n- Topic classification uses keyword matching (expect 40-50% \"Other\" for casual conversation titles)\n- Project attribution requires user-provided config and represents a lower bound\n- Cost metrics are designed for subscription pricing, not API pricing\n- Token counts are approximate and encoding-dependent\n\nThis repository contains no platform export files and distributes no conversation data. It exists to make long-horizon LLM usage measurable, reproducible, and inspectable without turning private conversations into a dataset.", "url": "https://wpnews.pro/news/open-source-toolkit-to-analyze-your-chatgpt-claude-usage-from-exports", "canonical_source": "https://github.com/noah-chelednik/llm-export-analytics", "published_at": "2026-05-27 05:33:55+00:00", "updated_at": "2026-05-27 05:57:15.940412+00:00", "lang": "en", "topics": ["large-language-models", "ai-tools", "generative-ai", "natural-language-processing", "ai-products"], "entities": ["ChatGPT", "Claude", "OpenAI", "GPT-4o", "GPT-5", "GPT-5.2-T", "Noah Chelednik"], "alternates": {"html": "https://wpnews.pro/news/open-source-toolkit-to-analyze-your-chatgpt-claude-usage-from-exports", "markdown": "https://wpnews.pro/news/open-source-toolkit-to-analyze-your-chatgpt-claude-usage-from-exports.md", "text": "https://wpnews.pro/news/open-source-toolkit-to-analyze-your-chatgpt-claude-usage-from-exports.txt", "jsonld": "https://wpnews.pro/news/open-source-toolkit-to-analyze-your-chatgpt-claude-usage-from-exports.jsonld"}}