A NotebookLM companion for batch downloading NSE company announcements
Project description
KnowledgeLM 🧠
A research companion for NotebookLM that automates the collection of NSE company filings. Designed to be operated by AI agents (Claude Code, Gemini CLI, etc.).
🚀 Instant Setup (AI Agents)
Install the knowledgelm-nse agent skill using:
npx skills add eggmasonvalue/knowledgelm-nse
This skill allows your AI agent to batch download a comprehensive bundle of investor materials(exchange filings - .pdfs and parsed XBRLs) and valuepickr forum thread for Indian publicly listed companies and optionally integrate them into NotebookLM for deep fundamental analysis.
✨ Features
- Agent-First: Optimized for LLMs with a standardized Agent Skill, and strictly formatted JSON output to
stdoutto preserve the context window. - Batch Downloads: NSE lacks bulk extraction; KnowledgeLM fetches filings by category in seconds via
fetch nse. - Granular XBRL Parsing: Extracts fine-grained facts (Personnel Changes/Resignations, Board Outcomes, Shareholder Meetings) into clean JSON using our standalone, highly optimized
nse-xbrl-parseroffline resolution engine. - Issue Documents: IPO prospectus, rights offers, QIP placements, information memoranda, and scheme documents — unified across 5 NSE endpoints.
- Markdown Conversion: Built-in
convertcommands to transform PDFs into LLM-ready Markdown. - NotebookLM Synergy: Purpose-built commands to facilitate source injection and bundled prompt templates for audio overviews.
- ValuePickr Forum Export: Export entire forum threads to clean, research-ready PDFs with reference extraction via
fetch vp. Besides, Valuepickr forum thread .pdfs are formatted for a distraction-free reading experience for humans. - Interactive UI: Browse and download individual filings (Resignations, Press Releases, etc.) via Streamlit.
🐧 ARM / Linux Setup
On ARM Linux devices, official Selenium Manager binaries are unavailable. You must manually install chromedriver using your system's package manager.
# Example (Debian/Ubuntu/Termux):
sudo apt install chromium-driver # Package names vary
Requirement: Ensure chromedriver is available in your $PATH.
🛠️ Manual Installation & Usage
CLI Usage
# Install from PyPI
uv tool install knowledgelm --upgrade
# Fetch corporate filings (NSE)
knowledgelm fetch nse HDFCBANK --start 2024-01-01 --end 2025-01-26 --datasets transcripts,annual_reports
# Fetch personnel changes/resignations (XBRL)
knowledgelm fetch nse RELIANCE --start 2024-01-01 --end 2025-01-26 --datasets personnel
# Fetch issue documents (IPO, rights, QIP, schemes)
knowledgelm fetch nse SWIGGY --start 2020-01-01 --end 2025-12-31 --datasets issue_documents
# Export forum thread (ValuePickr)
knowledgelm fetch vp "https://forum.valuepickr.com/t/hdfc-bank-limited/123" --symbol HDFCBANK
# Convert downloaded PDFs to Markdown
knowledgelm convert dir "./HDFCBANK_sources/transcripts/"
Web UI
Launch the interactive dashboard:
uv run streamlit run src/knowledgelm/app.py
📂 Project Structure
.agent/: AI Agent resources, skills, and prompt templates..context/: Living documentation (Architecture, Design, Changelog).src/: Core logic and Streamlit app.tests/: Unit and integration tests.
See .context/ARCHITECTURE.md for detailed diagrams.
📄 License
Distributed under the MIT License. See LICENSE for more information.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file knowledgelm-5.1.0.tar.gz.
File metadata
- Download URL: knowledgelm-5.1.0.tar.gz
- Upload date:
- Size: 153.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
982ddd0ffcfd6e115e8563f602acb85ecdb7d97612e0eb4f4ddd4d451ca46aae
|
|
| MD5 |
573d63b7ab21371843ee76b8522e966b
|
|
| BLAKE2b-256 |
a96fd31b3d1d7ec7ba587543b8b584e4ae1bdc426e0dd22829751dd74a6b03f4
|
File details
Details for the file knowledgelm-5.1.0-py3-none-any.whl.
File metadata
- Download URL: knowledgelm-5.1.0-py3-none-any.whl
- Upload date:
- Size: 38.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f199d1776c7bd638d316042e4271cc7ecb7e1268ad987c9be8631b20daa8416e
|
|
| MD5 |
4fb0772289b6d9e4c908a74f195e5057
|
|
| BLAKE2b-256 |
a401dfbd382dddebf3171c54e0c5b0eb8a82f4573dfd91d69d352719eb3ced06
|