ai-data-extractor vs Persefoni

Side-by-side comparison of features, pricing, and ratings

Analysis reviewed Live tool data as of 2026-09-29
Cross-checked through our multi-step verification ·
Saved

At a glance

Dimensionai-data-extractorPersefoni
What it isLocal Python CLI that flattens AI coding-assistant chat history into normalized JSONLAI-native carbon accounting platform for auditable Scope 1/2/3 emissions
Pricing modelFree (MIT license)Freemium, enterprise sales-led
BuyerIndividual developers, researchers, tinkerersSustainability teams at large enterprises and financial institutions
Coverage10 assistant sources (Claude Code, Codex CLI, Cursor, Windsurf, Trae, Continue, Gemini CLI, OpenCode, Cline, Roo Code, Aider)All 15 Scope 3 categories, Scope 1 and 2, plus PCAF financed emissions
Data locationLocal only — no account, no uploadCloud platform with Data Exchange and Integration Hub (AWS, Workiva, Patch)
AI layerNone; deterministic extractors plus heuristic parsersCopilot chat, Data Agent, Analytics Agent, Anomaly Detection, ISO/IEC 42001 certified
ai-data-extractor
ai-data-extractor

MIT-licensed Python CLI that pulls your AI coding assistants' local chat history into one normalized JSONL file.

Visit Website
Persefoni
Persefoni

AI-native carbon accounting software for audit-ready Scope 1, 2, and 3 emissions reporting

Visit Website
Pricing
Free
Freemium
Plans
—
$0/mo
Custom
Popularity
5 views
7.3k views
Skill Level
Intermediate
Intermediate
API Available
Platforms
CLIDesktop
WebAPI
Categories
🏷️ Data Labeling & Training Data💻 Code & Development📊 Data & Analytics
📜 GRC & Compliance Automation📊 Data & Analytics🧮 Business Intelligence
Features
Extract Claude Code session history from ~/.claude/projects/**/*.jsonl
Parse Codex CLI rollout files from ~/.codex/sessions/**/rollout-*.jsonl
Read Cursor chat data from SQLite state.vscdb in global and workspace storage
Heuristic extraction from Windsurf SQLite storage with an undocumented schema
Trae support combining SQLite and JSONL sources via heuristic parsing
Extract Continue sessions from ~/.continue/sessions/*.json
Extract Gemini CLI chats from ~/.gemini/tmp/<hash>/chats/*.json
Parse OpenCode session, message, and part trees
Extract Cline and Roo Code tasks storing raw Anthropic-format message arrays
Extract Aider markdown chat transcripts per project directory
Auto-detect macOS, Linux, and Windows data roots without OS flags
Normalize every source into one JSONL conversation-per-line format
Capture user messages, assistant responses, code context, diffs, and tool calls
CLI flags for --all, --sources, --list, --output-dir, --search-path, and --merge
Run extractors standalone with python -m extractors.<name> for debugging
Scope 1, 2, and 3 emissions measurement across all 15 Scope 3 categories
Market-based Scope 2 calculation for renewable energy tracking
PCAF-aligned financed emissions for loans and investments
Operational waste and water tracking by location, including hazardous waste
Persefoni Copilot: chat interface for technical carbon accounting questions
Anomaly Detection flags statistical outliers in datasets like electricity consumption
Natural-language emission factor mapping from spend files to LCA and Commodities factors (coming soon)
Persefoni Data Agent automates sustainability data preparation for reporting
Persefoni Analytics Agent enables conversational exploration of sustainability data
Sustainability Report Builder with guided workflows for CSRD, ISSB, and CDP
GHG Metrics Reports for CDP, CSRD, SECR, CA SB-253, and ISSB
GHG Methodology Report documents calculation methods and emission factor sets
Data Exchange for supplier, customer, and portfolio company data requests
Smart Forms, bulk uploads, APIs, and Integration Hub for emissions data ingestion
Inventory Management sets organizational boundaries, calculation methodology, and emission factor sets
Integrations
AWS
Workiva
Bain & Company
Patch
Connor Group
Amazon Sustainability Exchange

Feature-by-feature

The overlap here is essentially zero, so judge each on its own axis. ai-data-extractor's entire value is source coverage: ten assistant formats, including Claude Code's per-session JSONL under ~/.claude/projects, Codex CLI rollout files, Cursor's SQLite state.vscdb in both global and workspace storage, Continue's session JSON, Gemini CLI chats, OpenCode session/message/part trees, and Aider's per-project markdown transcripts, with Cline and Roo Code handling raw Anthropic-format message arrays one folder per task. Windsurf and Trae are supported only through heuristic parsers against undocumented schemas, which the project itself flags as not schema-stable. Output is one conversation per JSONL line, auto-detecting macOS, Linux, and Windows data roots.

Persefoni is a compliance system of record. It measures Scope 1, 2, and all 15 Scope 3 categories, adds PCAF-compliant financed emissions for loans and investments, and ingests activity data via Smart Forms, bulk uploads, APIs, and an Integration Hub. Its AI features — Copilot, the new Data Agent for automated data preparation, the Analytics Agent for conversational exploration, natural-language emission factor mapping, and anomaly detection — all exist to improve data quality before disclosure. Reporting is the product: Sustainability Report Builder for CSRD, ISSB, and CDP, GHG Metrics Reports covering CDP, CSRD, SECR, CA SB-253, and ISSB, plus a GHG Methodology Report documenting factor sets. Data Exchange pushes supplier and portfolio-company requests outbound.

Pricing compared

There is no pricing comparison to make. ai-data-extractor is free and MIT-licensed: you install a Python CLI, point it at your local app data, and get JSONL. There is no account, no seat count, no upload, and no vendor to pay — the cost is your own time reading Python extractors and adapting heuristic parsers for sources like Windsurf and Trae. Persefoni is freemium in name but enterprise in practice: the disclosure suite, PCAF financed emissions, Data Exchange, and the AI agents are sold through a sales conversation, and the product's own positioning says it is not for buyers who want to self-serve the full suite. Nor is it for small businesses that just need a cheap footprint calculator or non-regulated organizations with a single-scope footprint. Budget for the platform plus internal time, since Persefoni's AI layer depends on the quality of activity data you feed it. In short: one is a free script, the other is a procurement cycle.

Who should pick which

  • ML engineer building a fine-tuning corpus
    Pick: ai-data-extractor

    Normalizes ten assistant formats into one JSONL conversation-per-line file, locally.

  • Developer backing up chat logs
    Pick: ai-data-extractor

    Pulls Cursor's SQLite state.vscdb and Continue sessions to disk before an app clears them.

  • Sustainability lead at a CSRD filer
    Pick: Persefoni

    Report Builder, GHG Metrics Reports, and a methodology trail for CSRD, ISSB, and CDP.

  • Bank calculating financed emissions
    Pick: Persefoni

    PCAF-compliant accounting for loans and investments plus Data Exchange requests to portfolio companies.

  • Researcher comparing assistant session schemas
    Pick: ai-data-extractor

    The extractors double as documented examples of how each tool stores conversations.

Frequently Asked Questions

Could a company's sustainability team use ai-data-extractor for emissions data?

No. It reads AI coding-assistant chat history from local app data directories and emits JSONL conversations. It has no emissions factors, no disclosure templates, and no reporting outputs.

Is ai-data-extractor really free with no catch?

It is MIT-licensed and runs entirely locally with no account and no upload. The trade-off is support: there is no hosted service or SLA, and Windsurf and Trae parsing is heuristic against undocumented schemas, so those can break when those apps change.

Who is explicitly not a Persefoni customer?

Small businesses wanting a cheap simple calculator, non-regulated organizations with a single-scope footprint and no assurance requirement, teams wanting open-source or self-hosted accounting, self-serve buyers, and companies unwilling to invest internal time in data quality.

Does Persefoni's AI certification matter to a buyer?

It obtained ISO/IEC 42001 certification for its AI management system in August 2026, which is relevant if your procurement or assurance process requires documented AI governance. It says nothing about extraction tooling.

Which assistant sources does ai-data-extractor cover today?

Claude Code, Codex CLI, Cursor, Windsurf, Trae, Continue, Gemini CLI, OpenCode, Cline, Roo Code, and Aider — the last two added in the January 2026 releases.

Can I run Persefoni without a sales conversation?

The product's own not-for list includes buyers who need to self-serve the full disclosure suite, so expect a sales-led onboarding rather than instant checkout.

More ai-data-extractor or Persefoni comparisons

Explore each tool further

Browse these categories

Still deciding? Get the weekly AI tools brief

One email a week — new tools, honest comparisons, no spam.

Last reviewed: September 22, 2026