DailyInfo

中文 | English

DailyInfo is an automated research intelligence system for AI for Science researchers. It collects papers, AI news, code trends, and institutional updates, writes local Markdown briefings, pushes them to Discord, and now provides an agent-operated Zotero -> NotebookLM workflow for paper briefings and audio/video overviews.

Overview

Core flow:

FreshRSS / scrape / API sources
  -> dailyinfo run
  -> Markdown briefings
  -> dailyinfo push
  -> Discord channels + local archive

Agent-operated paper workflow:

Zotero today's additions
  -> dailyinfo zotero-brief
  -> PDFs + source_index.md
  -> NotebookLM briefing
  -> Audio Overview / Video Overview

Design principles:

Configuration-driven sources in config/sources.json.
Idempotent CLI commands that can be safely rerun.
External scheduling through cron, myopenclaw, openclaw, or other agent runtimes.
Clear capability/operator split: DailyInfo provides stable commands; Claude Code, Codex, or openclaw operate the workflow.

Screenshots

Put screenshots in pictures/ with these names and they will render here.

Discord Briefings

University Updates

Discord university updates briefing

Journal Papers

Discord journal papers briefing

arXiv Papers

Discord arXiv papers briefing

AI News

Discord AI news briefing

Discord code trending briefing

NotebookLM Audio Overview

NotebookLM audio overview

Data Layout

Default data root: ~/.myagentdata/dailyinfo/. Override it with DAILYINFO_DATA_ROOT.

~/.myagentdata/dailyinfo/
├── freshrss/data/       # FreshRSS SQLite + config
├── briefings/           # Markdown files waiting to be pushed
│   ├── papers/
│   ├── ai_news/
│   ├── code/
│   └── resource/
├── pushed/              # Successfully pushed archive
│   ├── papers/
│   ├── ai_news/
│   ├── code/
│   └── resource/
└── zotero/              # Zotero -> NotebookLM run packages
    └── YYYY-MM-DD[-collection]/
        ├── source_index.md
        ├── briefing_prompt.md
        ├── pdfs/
        ├── briefing.md
        ├── notebooklm.json
        └── MANUAL_NOTEBOOKLM_STEPS.md

Quick Start

git clone <repo-url>
cd dailyinfo

cp .env.example .env
# Fill OPENROUTER_API_KEY and DISCORD_BOT_TOKEN if you use the RSS/Discord pipeline.

uv sync --python python3
uv pip install -e .
dailyinfo install

# Optional: enable Zotero -> NotebookLM automation.
uv pip install -e ".[notebooklm]"

dailyinfo start
dailyinfo run
dailyinfo push

dailyinfo install validates .env, creates local data directories, and installs dependencies. It does not write crontab entries; scheduling belongs to your cron or agent runtime.

Main Commands

Command	Purpose
`dailyinfo install`	Validate environment and create data directories
`dailyinfo start` / `stop` / `restart`	Manage the FreshRSS container
`dailyinfo run`	Run all briefing pipelines
`dailyinfo run -p 1`	Pipeline 1: journal papers
`dailyinfo run -p 2`	Pipeline 2: AI news
`dailyinfo run -p 3`	Pipeline 3: arXiv CS.AI
`dailyinfo run -p 4`	Pipeline 4: code trending
`dailyinfo run -p 5`	Pipeline 5: university/resource
`dailyinfo run -f all`	Force regeneration for all sources
`dailyinfo push`	Push pending briefings to Discord and archive them
`dailyinfo push -d 2026-04-22`	Push briefings for a specific date
`dailyinfo status`	Show today's briefing/archive counts
`dailyinfo zotero-brief`	Prepare a Zotero -> NotebookLM paper briefing package
`dailyinfo zotero-brief --collection water --artifact audio`	Process the `water` collection and request Audio Overview
`dailyinfo zotero-brief --artifact video`	Request NotebookLM Video Overview
`dailyinfo zotero-brief --manual-only`	Prepare local materials without calling NotebookLM
`dailyinfo download-pdf <doi>`	PDF download via Claude Code (see docs)
`uv run python scripts/zotero_sync.py <pdf> <doi>`	Zotero linked_file sync (see docs)

Download PDF → Zotero Sync

Requires Claude Code + Playwright MCP plugin. For setup, supported publishers, and failure handling, see PDF Download and Zotero Sync.

Quick reference:

# One-time setup
npx playwright install chromium
uv pip install pyzotero

# Then in Claude Code:
/download-pdf 10.1038/s41586-026-10704-3

Zotero -> NotebookLM Agent Workflow

dailyinfo zotero-brief is a capability command, not the preferred daily user interface. The recommended daily interface is a local agent:

Claude Code slash command: .claude/commands/zotero-notebooklm.md
Codex skill: skills/zotero-notebooklm/SKILL.md
Future openclaw or other local runners can call the same CLI.

The workflow:

Reads Zotero papers by dateAdded.
Optionally restricts to a collection such as water.
Copies Zotero PDF attachments when available.
Opens Zotero attachment URIs to trigger cloud-drive hydration when PDFs are placeholders.
Uploads PDFs plus source_index.md to NotebookLM through notebooklm-py.
Asks NotebookLM to generate a Chinese paper briefing.
Optionally requests Audio Overview, Video Overview, or both.
Falls back to a local material package and manual steps if NotebookLM automation fails.

NotebookLM login is intentionally human-in-the-loop. The agent can open or prompt for the browser login, but the human completes Google authentication. The same NOTEBOOKLM_HOME must be used for login and later runs.

See:

Environment Variables

Variable	Purpose
`DEEPSEEK_API_KEY`	DeepSeek API key (primary model for `dailyinfo run`)
`DISCORD_BOT_TOKEN`	Discord bot token for `dailyinfo push`
`DISCORD_CHANNEL_PAPERS` / `_AI_NEWS` / `_CODE` / `_RESOURCE`	Optional category channel IDs
`FRESHRSS_USER`	FreshRSS username
`FRESHRSS_PASSWORD`	FreshRSS initial password
`DAILYINFO_DATA_ROOT`	Override default data root
`OPENROUTER_API_KEY`	OpenRouter API key (optional, used for fallback model)
`DAILYINFO_ENV`	Environment: `prod` / `dev` / `staging` (default `prod`)
`DAILYINFO_FALLBACK_MODEL`	Fallback model when DeepSeek returns empty (default `moonshotai/kimi-k2.5`)
`ZOTERO_LOCAL_BASE_URL`	Zotero local API base URL, default `http://127.0.0.1:23119`
`ZOTERO_API_KEY`	Zotero Web API key (for `zotero_sync.py` linked_file sync)
`ZOTERO_LIBRARY_ID`	Numeric Zotero user library ID (for `zotero_sync.py` linked_file sync)
`GDRIVE_PAPERS_PATH`	Local path to Google Drive papers folder (ZotMoov / Linked Attachment Base Directory)
`NOTEBOOKLM_HOME`	NotebookLM profile/auth directory used by `notebooklm-py`

Scheduling and Agents

DailyInfo intentionally avoids owning the scheduler. Recommended ownership:

Responsibility	Owner
FreshRSS container and local data layout	DailyInfo
Markdown generation via `dailyinfo run`	DailyInfo
Discord push/archive via `dailyinfo push`	DailyInfo
Timed execution	cron, myopenclaw, openclaw, or agent runtime
Zotero/NotebookLM orchestration	Claude Code, Codex skill, or local agent
Browser login and sensitive prompts	Human

Documentation

License

BSD 3-Clause License. See LICENSE for details.