Stop doomscrolling. Start knowing. FeedMind is a self-hosted, open-source AI-powered RSS briefing tool that reads all your feeds and delivers a clean, personalized 5-minute daily digest to your Telegram or email — every morning, zero noise.
You follow 50+ sources — security blogs, CVE feeds, GitHub releases, newsletters, tech news. You can't read them all. You shouldn't have to.
FeedMind reads everything for you. You get the summary.
Every morning at 7:00 AM:
🛡️ FeedMind Daily Briefing — May 9, 2026
🔴 CVE-2026-12345 (CVSS 9.8): RCE in Apache Log4j2 — Patch NOW
🟡 Cisco ASA Advisory: Privilege escalation — patch available
📌 Krebs on Security: New EDR evasion technique in the wild
🐙 Trivy v0.45.0 released — improved SBOM support for containers
🐙 Vault v1.16.1 — patches CVE-2026-XXXX in token renewal
→ 5 sources · 4 min read
No apps to open. No tabs to check. Just what matters, in your Telegram.
- 📡 Reads any RSS/Atom feed — blogs, CVE databases, GitHub releases, newsletters
- 🧠 AI-powered summarization — uses a local LLM (Ollama) or cheap API (GPT-4o mini)
- 📱 Telegram delivery — briefing arrives in your chat every morning
- 🔒 Privacy-first — everything runs on your server; your feeds never leave your machine
- 🐳 Docker-ready — up and running in under 60 seconds
- 🛡️ Security mode built-in — pre-configured feeds for CVEs, CISA advisories & threat intel
- 💸 Near-zero cost — RSS is plain text; local LLMs cost fractions of a cent per summary
- 🔓 Open-source — MIT license, no vendor lock-in, yours forever
- Docker & Docker Compose
- A Telegram bot token (get one from @BotFather)
- Your Telegram chat ID (get it from @userinfobot)
- Ollama running locally (install Ollama) or an OpenAI API key
git clone https://github.com/black-hak/feedmind.git
cd feedmindcp .env.example .env
cp feeds.txt.example feeds.txtEdit .env with your Telegram token, chat ID, and LLM settings.
Edit feeds.txt to add your RSS sources (one URL per line).
docker-compose up -dThat's it. FeedMind will deliver your first briefing at the time you configured.
FeedMind ships with a pre-configured security bundle — perfect for security professionals, SOC teams, and developers who need to stay on top of threats without drowning in noise.
Just use feeds.txt.example as-is and you'll get daily briefings covering:
| Source | What you get |
|---|---|
| NVD | All new CVEs, filtered by CVSS score |
| CISA Advisories | Actionable alerts from US Cyber Defense |
| Krebs on Security | Trusted threat intel & deep analysis |
| Bleeping Computer | Breaking security news |
| Exploit-DB | Active exploits published today |
| Packet Storm | Vulnerability disclosures & PoCs |
| GitHub Releases | Updates for tools you depend on (Trivy, Vault, etc.) |
The default LLM prompt is tuned to highlight only what requires action: critical CVEs (CVSS ≥ 9.0), active exploits, and advisories with patches available.
# ─── Telegram ──────────────────────────────────────────
TELEGRAM_BOT_TOKEN=your_bot_token_here
TELEGRAM_CHAT_ID=your_chat_id_here
# ─── LLM Provider ──────────────────────────────────────
# Options: ollama (local, free) | openai (cloud, cheap)
LLM_PROVIDER=ollama
LLM_BASE_URL=http://localhost:11434
LLM_MODEL=llama3:8b
# ─── AI Prompt ─────────────────────────────────────────
# Customize what the LLM focuses on.
# Default is tuned for security teams.
LLM_PROMPT="You are a senior security analyst. From the following feed items, extract ONLY: 1) Critical vulnerabilities (CVSS ≥ 9.0) published in the last 24h, 2) Advisories requiring immediate action (patch available), 3) Active exploits published today, 4) Notable tool releases from trusted sources. Use 🔴 for critical, 🟡 for medium, 🐙 for releases. Be concise. Max 600 characters. Ignore marketing, events, and non-actionable content."
# ─── Scheduling ────────────────────────────────────────
FEEDS_FILE=./feeds.txt
DAEMON_HOUR=7 # Delivery hour (0–23, local time)
FETCH_INTERVAL=1h # How often to poll feeds in background
SQLITE_DB=./data/feedmind.dbOne URL per line. Lines starting with # are comments.
# Security feeds
https://nvd.nist.gov/vuln/data-feeds/JSON_FEED
https://www.cisa.gov/cybersecurity-advisories/feed
https://krebsonsecurity.com/feed/
# GitHub releases (any repo you care about)
https://github.com/aquasecurity/trivy/releases.atom
https://github.com/hashicorp/vault/releases.atom
# Tech blogs
https://news.ycombinator.com/rss
feedmind/
├── cmd/ # CLI entrypoint
│ └── main.go
├── internal/
│ ├── feeds/ # RSS/Atom parser & storage
│ │ └── parser.go
│ ├── llm/ # LLM integration (Ollama / OpenAI)
│ │ └── summarizer.go
│ ├── notify/ # Delivery channels (Telegram, email)
│ │ └── telegram.go
│ └── scheduler/ # Cron-based daemon
│ └── daemon.go
├── data/ # SQLite DB (auto-created, gitignored)
├── feeds.txt.example # Pre-configured security feeds
├── .env.example # Configuration template
├── Dockerfile
├── docker-compose.yml
go.mod
go.sum
FeedMind is designed so your data never has to leave your machine:
- All feed data is stored in a local SQLite database (or your own Postgres instance)
- When using Ollama, LLM inference runs 100% locally — no data sent to any API
- Even when using OpenAI/Anthropic, only the feed summaries (not raw content) are sent
- No telemetry, no analytics, no accounts required for self-hosted mode
- Core RSS/Atom parser
- LLM summarization (Ollama + OpenAI)
- Telegram delivery
- Docker + self-hosted support
- Email delivery (SMTP)
- Interactive Telegram bot (
/addfeed,/setprompt,/brief) - Web UI for feed management
- Cloud-hosted version — feedmind.app ($8/month, no server needed)
- B2B security teams plan ($49/month, shared briefings + Slack integration)
- Voice briefing via TTS (listen while you make coffee)
Pull requests are welcome. For major changes, please open an issue first.
- Fork the repo
- Create your branch:
git checkout -b feat/your-feature - Commit your changes:
git commit -m 'feat: add your feature' - Push to the branch:
git push origin feat/your-feature - Open a Pull Request
MIT — free to use, modify, and distribute.
Built with ☕ and 🧠 for the people who need to know everything but have time for nothing.