PoC-Version-Final-Snapshot

This is a snapshot release of the final gradio PoC build.
This version was the first iteration of the application, and is no longer supported.
If you are looking for a replacement for a local-first/open-source solution similar, please see either tldw_chatbook: https://github.com/rmusser01/tldw_chatbook or the Web Front-end for tldw_server:

Install instructions:

tldw_server – Installation Guide (Win/Mac/Linux)

Requirements

• Python 3.9+
• ffmpeg (media processing)
• 8GB+ RAM (3–4GB for server, rest for models)
• 10GB+ disk space

---

1. Install System Dependencies

Linux

# Debian/Ubuntu
sudo apt install ffmpeg portaudio19-dev gcc build-essential python3-dev
# Fedora
sudo dnf install ffmpeg portaudio-devel gcc gcc-c++ python3-devel

macOS

brew install ffmpeg portaudio

Windows

• Install Python and ffmpeg from official sources.
• For CUDA transcription without full CUDA install:
  Download [Faster-Whisper-XXL](https://github.com/Purfview/whisper-standalone-win/releases/download/Faster-Whisper-XXL/Faster-Whisper-XXL_r192.3.4_windows.7z)
  Extract cudnn_ops_infer64_8.dll and cudnn_cnn_infer64_8.dll into the tldw_server directory.
  Download/install ffmpeg from https://www.gyan.dev/ffmpeg/builds/ - Move the ffmpeg/ffmprobe binaries into the ./bin folder if you don't install them system-wide. This will let tldw use ffmpeg for transcription/file conversion.

---

2. Download Software & Create Virtual Environment

# Download this package via your choice, and then extract the folder to a location of your choosing, and then navigate to that folder via the terminal.

cd tldw_server

python3 -m venv .venv
# Activate:
# Linux/macOS:
source .venv/bin/activate
# Windows:
.venv\Scripts\activate

---

3. Install Python Dependencies

pip install -r requirements.txt

---

4. Configure API Keys

cp config.txt.example config.txt
# Edit config.txt with your keys & settings

Alternatively, use environment variables (OPENAI_API_KEY, ANTHROPIC_API_KEY, etc.).

---

5. Run the Server

python summarize.py -gui

---

Optional – Docker Install

CPU

docker build -f Helper_Scripts/Dockerfiles/tldw_cpu_Dockerfile -t tldw-cpu .
docker run -p 8000:8000 tldw-cpu

GPU

docker build -f Helper_Scripts/Dockerfiles/tldw_nvidia_Dockerfile -t tldw-gpu .
docker run --gpus all -p 8000:8000 tldw-gpu

PoC-Version-Final-Snapshot

This release is the final snapshot of the Proof-of-Concept version of tldw.

It was built using the original tldw project as a donor base and extended from there, using Gradio as the UI.

This project is meant for a single user and while it can definitely be useful, I would recommend looking towards the latest release of tldw for a better codebase, and less bugs.

What's Changed

• Multi-file audio upload by @rmusser01 in #437
• Re-commented out function profiling and 'fixed' The multi-file audio upload by @rmusser01 in #438
• Adding Streaming for replies by @rmusser01 in #449
• Docs by @rmusser01 in #452
• Added Docs for current/planned features by @rmusser01 in #455
• Docs, TTS, answer streaming, web search and perplexity clone by @rmusser01 in #457
• config.txt Replaced by @rmusser01 in #472
• Fixes for api Call Streaming checks (one more fix on the way for minp/topk values for APIs) by @rmusser01 in #473
• Wew, i'm an idiot. by @rmusser01 in #476
• Add Security fix notice by @rmusser01 in #477
• Fixes for commercial APIs + Adding Top-P / Min-P / Top-K as API option fields (even if they aren't exposed in the UI yet...) by @rmusser01 in #479
• Kokoro TTS + Whatever makes it in until I merge this by @rmusser01 in #490
• TTS Bugfixes for Kokoro by @rmusser01 in #509
• Another TTS fix by @rmusser01 in #512
• Config loading fixes by @rmusser01 in #513
• token count for Kokoro by @rmusser01 in #514
• Chat fixes + DL generated TTS from chats by @rmusser01 in #515
• Pyproject.toml file added/Initial support for optional features ; Unit tests for character card library ; preview box for plain text/markdown/docx/rtf files ; Fixes for epub parsing by @rmusser01 in #517
• Alltalkbugfix; Ooba/vllm working; switched to Loguru and hash checks on ingestion by @rmusser01 in #519
• Update config.txt by @rmusser01 in #535
• Update Backup_Config.txt & Fix Cohere Summarization by @rmusser01 in #536
• Update Summarization_General_Lib.py by @rmusser01 in #538
• Max response tokens setting, bugfix for custom openAI API endpoint by @rmusser01 in #539
• Fix for llama.cpp summarization by @rmusser01 in #542
• Fix for llama.cpp chat API + local timeout value by @rmusser01 in #547
• Update Utils.py by @rmusser01 in #548
• Fix ollama + loosen validity check for printable characters in summary results by @rmusser01 in #549
• Update Utils.py by @rmusser01 in #553
• Additional logging for ollama timeouts by @rmusser01 in #555
• Links + More Progress by @rmusser01 in #556
• Forgot to include the 'config.txt' changes... by @rmusser01 in #564
• Fixes + Links + Update to README by @rmusser01 in #565
• Update README.md by @rmusser01 in #568

Full Changelog: Beta-v0.8...PoC-Version-1.0

Beta-v0.8-SNAPSHOT

What's Changed

• Potential fix for #388 by @rmusser01 in #389
• XML Ingestion + Default global API by @rmusser01 in #396
• Anki, install script bugfix, README update, Bookmarks->URLs for scraping, new view tab for Media/RAG notes by @rmusser01 in #397
• Update README.md by @rmusser01 in #398
• Fixed RAG Note saving + Transcript saving when the existing file is blank by @rmusser01 in #399
• Semantic Scholar integration by @rmusser01 in #400
• Mindmap creation via PlantUML (Utilties tab) + Token counts in conversations by @rmusser01 in #405
• Removed Chats from Media DB (BREAKING CHANGE - moved to RAG QA db), Added mindmap viewing, token counts in convos, anki flashcard validation by @rmusser01 in #409
• SQLite MediaDB metrics by @rmusser01 in #413
• DB Views Fix + new run through video by @rmusser01 in #414
• Update README.md by @rmusser01 in #415
• Update Anki_tab.py by @rmusser01 in #416
• Update Windows_Install_Update.bat by @rmusser01 in #417
• More prompt fixes by @rmusser01 in #418
• Better PDF ingestion/parsing via Docling + pymupdf4llm by @rmusser01 in #419
• Updated README by @rmusser01 in #420
• Fix for plain RAG chat (refactored 'enhanced_rag_pipeline') by @rmusser01 in #421
• RAG fix for blank results by @rmusser01 in #422
• Update PDF_ingestion_tab.py by @rmusser01 in #424
• Add Local API timeout option in config.txt by @rmusser01 in #426
• Added Google API + Support for multiple video upload by @rmusser01 in #429
• Eval Plans writeup by @rmusser01 in #433

Full Changelog: Beta-v0.7...Beta-v0.8

Beta v0.7 Release

WIP v0.7 Release - SNAPSHOT
Differences from the last one:
idk.
RAG Notes, thats all that comes to mind.
Obviously much more, but eh.

This is meant only for archival/snapshot-in-time purposes. If you are looking to use the project, clone the latest from MAIN.

Beta v0.6

Snapshot

Beta v0.5 release

Code Snapshot

What's Changed

Stuff
RAG

Full Changelog: Beta-v0.4...Beta-v0.5

Beta v0.4 release

Snapshot of latest checkpoint.

What's Changed

• Dev by @rmusser01 in #138
• Added some more features by @rmusser01 in #139
• Chat tabs + prompt keywords by @rmusser01 in #146
• Dev by @rmusser01 in #147
• Merge fixes by @rmusser01 in #148
• Summary Review/Search is in by @rmusser01 in #149
• fixups to tabby and vllm. Still haven't tested them. by @rmusser01 in #150
• Dev by @rmusser01 in #151
• Dev by @rmusser01 in #152
• Dev by @rmusser01 in #153
• Update README.md by @rmusser01 in #154
• Update README.md by @rmusser01 in #155
• Fix system prompts + add 'system_message' arg to API calls by @rmusser01 in #156
• Added system prompt inputs to various functions. by @rmusser01 in #157
• Dev by @rmusser01 in #163
• Ollama api support by @rmusser01 in #165
• Removed 'editable' chat tab + Summary Comparison page by @rmusser01 in #166
• Fixed chat saving + prompt typo. by @rmusser01 in #168
• PDF ingestion works. by @rmusser01 in #172
• Sync latest changes by @rmusser01 in #173

Full Changelog: Beta-v0.2...Beta-v0.4

Beta V0.2

This is a snapshot release for archival purposes.
This release should work as expected.