AutoRename-PDF v3.2.0

AutoRename-PDF v3.2.0

Windows GUI now matches the CLI: AP/AR defaults, gpt-5.6-luna, and a shorter example-led README. Recipe configs live on GitHub (examples/), not in the portable ZIP.

GUI

• Visual system and settings closeout (titlebar, files, status bar)
• Sidecar spawn so Preview / Rename / Undo talk to the signed CLI
• Native-app README screenshots (not Vite)

Docs and examples

• config.yaml.example is the one-stop copy-me file
• Recipes: academic, ER/AR, invoice-id, Ollama offline, vision-cloud
• Portable ZIP stays flat: EXE pair, setup.ps1, and example templates only

Testing

• 637 unit tests passing (54 live-provider tests skipped by default)
• GUI: typecheck + 60 Vitest tests

AutoRename-PDF v3.1.0

AutoRename-PDF v3.1.0

Default business names now look like 20260906 ACME AP 12,13.pdf. Existing v2 configs still load; generated names and prompts change on purpose.

Dynamic extraction profiles

• Built-in business and academic profiles, plus YAML overlays, inheritance, and rename --profile
• Business filenames add the printed invoice total as the fourth component; missing totals omit it (20260906 ACME AP.pdf)
• Invoice type codes default to AP / AR (Accounts Payable / Accounts Receivable). Restore German DATEV-style codes with pdf.incoming_invoice: ER and pdf.outgoing_invoice: AR
• JSON output includes profile and raw fields; config validate / config show cover resolved profiles
• Opt out of the fourth filename token with a three-component business template (see README)

Stack and extraction

• OpenAI and Anthropic use native structured parse; Instructor remains for Gemini, xAI, and Ollama
• Default OpenAI example model is gpt-5.6-luna
• PaddleOCR embed upgraded to PP-OCRv6 small (PaddleOCR 3.7.0 / PaddlePaddle 3.3.1, Python 3.13.15)
• Rename pipeline accepts OCR-only content; setup elevation is limited to Explorer context-menu registry actions
• GUI: Vite 8, current Tauri 2 plugins, pnpm, Vitest 5; frozen lockfile install before Tauri builds
• Windows CI runs ruff + pytest

README and setup

• README leads with invoice, Explorer, offline, academic, and CLI paths
• setup.ps1 keeps config and PaddleOCR under the original Windows user; only context-menu install/remove requests UAC

Migration

• If an old prompt_extension appends the amount to document_type, remove it or amounts can appear twice
• Configs that already set incoming_invoice: ER keep German codes; omitted keys pick up AP

Testing

• 605 unit tests passing (54 live-provider tests skipped by default)

AutoRename-PDF v3.0.1

AutoRename-PDF v3.0.1

PaddleOCR 3.4.0 Compatibility

• Fixed: NameError: name 'en' is not defined during setup — PowerShell string quoting issue
• Fixed: ValueError: Unknown argument: show_log — removed deprecated parameter
• Fixed: Wrong recognition model loaded — now uses explicit model mapping per language/script family
• Fixed: PermissionError on model files installed by admin — setup.ps1 now sets ACLs via locale-independent SID
• Fixed: Excessive RAM usage — switched from PP-OCRv5_server to PP-OCRv5_mobile models

Config Changes

• Simplified: paddleocr.languages (list) → paddleocr.language (string) — old config auto-migrates
• Language codes now map to script-family models: en → English-only, de/fr/es/etc. → Latin (covers all 59 Latin-script languages including English), ch → Chinese+Japanese
• See config.yaml.example for full documentation

Improvements

• Real-time stderr streaming during PaddleOCR model downloads — no more silent waits
• Setup.ps1: UAC elevation message, better error handling for model preload
• GUI: Language setting is now a simple text field with hint

Testing

• 354 tests passing (unit + live integration across OpenAI, Anthropic, Ollama)

v3.0.0

v3.0.0

Major Release: GUI, Multi-Provider AI, and Three-Tier Extraction

AutoRename-PDF is now a full desktop application with a Tauri-based GUI, support for 5 AI providers, and a three-tier extraction pipeline.

What's New

• Desktop GUI — Native Windows app built with Tauri 2, featuring drag-and-drop, dark theme, and real-time rename previews
• Multi-provider AI — Choose between OpenAI, Anthropic, Google Gemini, xAI (Grok), or local Ollama models
• Three-tier extraction — Text extraction (pdfplumber), optional OCR (PaddleOCR), and optional vision mode (send page images to LLM) for scanned/image-heavy PDFs
• Structured output — All providers use instructor + Pydantic for reliable, typed AI responses
• CLI JSON mode — --output json for scripting and automation
• Environment variable support — Use ${VAR_NAME} syntax in config.yaml, with .env file auto-loading
• Comprehensive test suite — 344 tests, 95% coverage, including live integration tests across providers

Breaking Changes

• Config schema v2 — Old v1 configs (openai.api_key) must be migrated to the new flat ai.* structure. See config.yaml.example.
• use_gpu renamed to device — PaddleOCR config now uses device: "auto" | "cpu" | "gpu" instead of use_gpu: true/false. Old configs are auto-migrated with a deprecation warning.
• Python 3.11+ required

Bug Fixes

• Fixed Anthropic vision mode sending OpenAI-format image_url blocks instead of native image blocks
• Updated default and test models to current generation (GPT-5.4, Claude Haiku 4.5)

Distribution

• Portable ZIP: GUI + CLI + setup script
• Context menu integration via setup.ps1
• Optional PaddleOCR install (~500MB) for free local OCR

v2.1.2

What's Changed

• Add funding section to README with business value messaging by @Copilot in #10
• Fixed setup.ps1

New Contributors

• @Copilot made their first contribution in #10

Full Changelog: v2.1.1...v2.1.2

v2.1.1

Full Changelog: v2.1.0...v2.1.1

Fixed --dpi bug (halucinated parameter)

v2.1.0

What's Changed

• Refactored to yaml for easier use (.env has some problems with cloud sync and yaml is more end user friendly)
• NOTE: Use an online tool to convert json to yaml, eg: https://jsonformatter.org/json-to-yaml
• added Private GPT ability by @FrankBie in #5
• IMPORTANT; We will not allow any Pull requests from now on, please use Issues

New Contributors

• @FrankBie made their first contribution in #5

Full Changelog: v2.0.1...v2.1.0

v2.0.1

Full Changelog: v2.0.0...v2.0.1

Much more internalization

How to update:

• download the .zip file and extract all files over your previous installation
• open the .env file
• copy all new .env variables from .env.example to your .env file and adapt as needed
• if you added/changed your OCR_LANGUAGES, re-run the setup.ps1 in administrator mode.

v2.0.0

Full Changelog: v1.1.2...v2.0.0

No more poppler, switched to ocrmypdf and pymupdf.

Way easier installation.

Please consider funding 🙏

For Installation see README

v1.1.2

Full Changelog: v1.1.1...v1.1.2