fleet-wide distributed honeypot mesh built on Hyperswarm and HyperDB
# Website: https://honeypeer.com
Today I am announcing the **public Alpha** of [**HoneyPeer**](https://honeypeer.com) — operated by **HoneyPeer, LLC**, a Georgia limited liability company formed in DeKalb County. After years of R&D and a full year of building, testing, and hardening a peer-to-peer honeypot platform from the ground up, I am inviting operators, homelab enthusiasts, and security-minded teams to join the mesh **at no cost** while I finish the product for general availability. This is not a slide-deck honeypot. HoneyPeer is a **running system**: subscription-gated peers, a coordinator that ingests real attack telemetry, a detection engine that promotes IPs through reputation tiers, and **fleet-wide kernel firewall sync** so one probe anywhere can protect everyone. If you have ever wanted collective immunity against internet background radiation — SSH brute force, Redis scans, SMTP probes, credential stuffing against fake admin panels — this is what I built it for. ## The problem I set out to solve Every server on the public internet is scanned constantly. Most of that traffic is automated: botnets mapping open ports, exploit kits checking for default credentials, and opportunistic attackers looking for the next foothold. Traditional defenses — perimeter firewalls, WAFs, rate limits — help, but they are **local**. Your blocklist does not automatically become your neighbor's blocklist. Your honeypot findings sit in a CSV until someone manually imports them. HoneyPeer flips the model: 1. **Decoys attract** high-signal probes on real protocol handlers (SSH, HTTP, SMTP, databases, and more). 2. **Peers report** structured telemetry upstream over encrypted P2P transport. 3. **The coordinator scores** source IPs with sliding windows, reputation tiers, and subnet co-conspirator detection. 4. **Blocks replicate** through HyperDB to every licensed node and apply in the **kernel** (nftables, ipset, iptables, or macOS `pf`). One attack seen anywhere can protect the entire fleet — often in under a second from report to fanout. ```mermaid flowchart LR subgraph internet [Internet scanners] A[Attacker] end subgraph peer [Your peer] H[Honeypots] FW[Kernel firewall] P[Embedded panel] end subgraph mesh [HoneyPeer mesh] M[Coordinator] DET[Detection engine] DB[(HyperDB)] end A --> H H -->|attack report| M M --> DET DET -->|block tier| DB DB -->|replicate| FW M -->|banIp fanout| FW P --> H ``` ## What HoneyPeer is (and what it is not) **HoneyPeer is:** - A **distributed honeypot SaaS** where each subscriber runs one or more **peer nodes** on infrastructure they control. - A **Hyperswarm mesh** connecting licensed peers to a coordinator (master) for registration, attack ingest, block snapshots, and database replication. - A **Bare-runtime peer binary** (single-file, cross-platform) with 25+ protocol-accurate honeypot listeners and an embedded operator panel. - A **privacy-conscious** platform: marketing pages do not load third-party analytics; sensitive billing data stays in a master-local database that **never replicates** to peers. **HoneyPeer is not:** - A managed cloud honeypot you point DNS at (you run the peer). - A replacement for your production WAF or IDS (it complements them with **shared threat signal**). - A guarantee of zero false positives (tune detection via env vars; whitelist trusted IPs on each peer). ## Architecture at a glance HoneyPeer-NG splits cleanly into three layers: | Layer | Component | Role | |-------|-----------|------| | **Edge** | Peer honeypots | Listen on curated TCP/UDP ports; log interactions; optional tarpit | | **Transport** | Hyperswarm + Protomux | Encrypted P2P: register, report attacks, replicate HyperDB, push `banIp` | | **Control** | Master coordinator | Writable network DB, detection engine, Stripe licensing, admin UI | ### Two databases by design Security and billing requirements forced a hard split: | Database | Replicates to peers? | Contents | |----------|----------------------|----------| | **Network-replicated** | Yes | Attacks, IP blocks, peer blocks, peers, subscription metadata (non-sensitive) | | **Local-customer** | **Never** | Sales, license keypairs, customer emails, Stripe IDs | Your peer receives enough data to enforce blocks and prove subscription status. It never receives another customer's keypair or payment details. ### Ports you will interact with | Surface | Default port | Purpose | |---------|--------------|---------| | Peer panel | 8888 | Local honeypot + firewall control on each node | ## Honeypots: high-signal decoys, not generic banners Each peer ships a **registry of protocol handlers** implemented on the [Bare](https://github.com/holepunchto/bare) runtime — the same stack powering modern Holepunch peer infrastructure. Handlers speak enough of the real protocol to keep scanners engaged and produce rich `interactionData` for the detection pipeline. Default known slots include: | Port | Service | Port | Service | |------|---------|------|---------| | 22 | SSH | 443 | HTTPS | | 23 | Telnet | 445 | SMB banner | | 25 | SMTP | 1433 | MSSQL | | 53 | DNS (UDP) | 1521 | Oracle TNS | | 80 / 8080 | HTTP | 3306 | MySQL | | 161 | SNMP | 3389 | RDP | | 389 | LDAP | 5432 | PostgreSQL | | — | — | 6379 | Redis (interactive RESP) | | — | — | 27017 | MongoDB wire | | — | — | 6443 | Kubernetes API | | — | — | 9200 | Elasticsearch | Peers also spin up **random high ports** with Iighted novel handlers and optional **tarpit** mode to slow scanners. Deception profiles, canary credentials, and HTTP personality switches let you tune how aggressive the fiction is — without npm protocol libraries on the hot path. ## From probe to fleet-wide block Here is the attack lifecycle in plain language: 1. **Probe** — An IP connects to e.g. SSH :22 on your peer. The handler accepts, logs metadata, and may record credential attempts. 2. **Report** — The peer packages a structured attack record and sends it to the coordinator over the swarm link. 3. **Score** — Detection windows update IP reputation. Greylist tiers accumulate signal; **block** tier persists a global ban and triggers fanout. 4. **Replicate** — HyperDB pushes the block to every licensed peer's read-only replica. 5. **Enforce** — Each peer applies the ban in its kernel firewall backend. Whitelist entries always win. Subnet **co-conspirator** logic groups activity from the same /24, so coordinated scan waves surface as related clusters in the dashboard — not isolated one-off events. ```mermaid sequenceDiagram participant S as Scanner participant P as Peer honeypot participant M as Coordinator participant D as Detection participant F as Fleet peers S->>P: TCP probe / login attempt P->>M: attack-report (Protomux) M->>D: ingest + score IP D->>M: promote to block tier M->>F: banIp + HyperDB replicate F->>F: nftables / pf apply ``` ## Alpha program: free fleet access while I ship **Alpha is open now.** I want real networks, real noise, and real feedback — not lab-only simulations. | | Alpha (today) | Launch (forecast) | |---|---------------|-------------------| | **Price** | **$0.00/month** — any seat count 1–50 | From **$4.99/node/month** with volume discounts | | **Features** | Full mesh: honeypots, detection, blocks, panel, viz | Same core platform | | **Checkout** | Stripe **Test mode** (sandbox) | Stripe Live | | **Keypairs** | One Noise keypair per seat, provisioned at checkout | Same licensing model | ### What Alpha checkout does 1. You choose **single-node** or **multi-node** (2–50 peers) on the [pricing page](https://honeypeer.com/pricing). 2. Stripe Hosted Checkout runs in **test mode** — use test card patterns only; do not enter real payment details during Alpha. 3. On success you receive a **subscription record** and downloadable **peer keypair(s)** — one per seat. 4. You download the peer binary for your platform and connect to the mesh. **Important:** Sandbox subscriptions **do not carry over** to Stripe Live at launch. When paid billing begins, you will complete a **fresh checkout on the live network**. I will announce pricing in advance and give Alpha participants reasonable notice before any charge. Your Alpha keypairs remain valid during the free period. Legal details: [Terms of Service — Alpha](https://honeypeer.com/legal/terms#alpha) · [Privacy Policy](https://honeypeer.com/legal/privacy) · [Peer EULA](https://honeypeer.com/legal/peer-eula). ## Getting started in four steps ```mermaid flowchart TD A[Visit honeypeer.com/pricing] --> B[Alpha checkout — pick seats] B --> C[Download keypair + peer binary] C --> D[Run peer with keypair] D --> E[Open local panel :8888] E --> F[Watch attacks on viz.honeypeer.com] ``` ### Step 1 — Claim your Alpha license Go to **[honeypeer.com/pricing](https://honeypeer.com/pricing)** and start checkout. Multi-node subscriptions provision **one keypair per seat** so each peer has a distinct identity in the mesh. ### Step 2 — Download the peer Prebuilt single-file binaries ship for: | Platform | Artifact | |----------|----------| | Linux x64 / arm64 | `honeypeer-peer-linux-*` | | macOS (universal) | `honeypeer-peer-macos-universal` | | Windows x64 / arm64 | `honeypeer-peer-windows-*.exe` | Downloads: **[honeypeer.com/download](https://honeypeer.com/download)** and the [latest release](https://git.ssh.surf/snxraven/honeypeer-releases/releases/tag/latest) mirror. ### Step 3 — Run the peer ```bash chmod +x honeypeer-peer-linux-x64 ./honeypeer-peer-linux-x64 --keypair /path/to/your-keypair.json ``` Use **sudo** (or `CAP_NET_ADMIN` + `CAP_NET_RAW`) when you want kernel firewall enforcement. The embedded panel defaults to **http://localhost:8888**. Peer configuration reference: [docs/peer/installation](https://docs.honeypeer.com/peer/installation) (operator docs). ### Step 4 — Observe the mesh - **Live network map:** [viz.honeypeer.com](https://viz.honeypeer.com) — public attack activity (IPs cloaked for privacy). - **Threat blog:** AI-assisted briefings on patterns I observe fleet-wide. - **Discord:** [join.honeypeer.com](https://join.honeypeer.com) — community support and operator chat. ## Operator surfaces beyond the peer panel HoneyPeer is a full operator platform, not just a background daemon. ### Customer support portal (`support.honeypeer.com`) Subscribers can open tickets, browse docs, and manage their relationship with the operator without accessing the admin dashboard. ### Documentation (`docs.honeypeer.com`) Installation, peer firewall backends, honeypot handler matrix, API reference, and architecture deep-dives for anyone self-hosting or auditing the stack. ## Built on peer-to-peer infrastructure HoneyPeer stands on the Holepunch ecosystem: - **Hyperswarm** for topic discovery and NAT-friendly connections - **HyperDB / Hypercore** for replicated, append-only network state - **Bare** for a lightIight peer runtime without Node.js on the edge - **Protomux** for typed channels betIen coordinator and peers I believe defensive tooling should inherit the same resilience properties as the open internet: no single vendor choke point, encrypted transport by default, and peers that can enforce policy even when they are only loosely coupled to the coordinator. ## Security and privacy commitments - **Subscription-gated mesh** — only licensed keypairs register; peer public-key blocks can eject abusive nodes at the Hyperswarm firewall layer. - **Split data plane** — billing secrets never replicate. - **No third-party analytics** on marketing pages. - **Whitelist wins** — your trusted IPs are never kernel-blocked by global bans. - **Backups** — scheduled tar archives of network + local DB for disaster recovery (operator-configured). Read more: [Security & privacy](platform/security-and-privacy.md) · [Detection & blocks](platform/detection-and-blocks.md). ## What I am looking for during Alpha I am especially interested in feedback on: - **Detection tuning** — false positives, missed scans, tier promotion timing - **Firewall backends** — nftables vs ipset vs pf on your OS images - **Honeypot fidelity** — protocols that attract noise but lack handler depth - **Fleet scale** — replication lag and fanout behavior past dozens of peers - **Operator UX** — dashboard workflows, docs gaps, checkout confusion File issues, open support tickets, or talk to us on Discord. This mesh gets better when real operators stress it. ## Roadmap preview (post-Alpha) I am not promising dates — but the direction is clear: - **Paid launch** on Stripe Live with transparent per-node pricing and volume tiers - **Continued protocol expansion** and richer interaction capture - **Deeper viz and reporting** for fleet owners - **Enterprise options** — dedicated coordinators, SLAs, custom retention (inquiry via support) Launch pricing forecast: **$4.99/node/month** list, with automatic fleet discounts up to **25% off** at 20–50 nodes. The [pricing calculator](https://honeypeer.com/pricing) shows live forecasts. ## A note on the company **HoneyPeer, LLC** is a Georgia limited liability company formed in **DeKalb County, Georgia, United States**. I build and operate the HoneyPeer network as the licensor under our [Peer End User License Agreement](https://honeypeer.com/legal/peer-eula). Copyright © 2026 HoneyPeer, LLC. ## Join the Alpha today If you operate servers, a homelab, or a small fleet that touches the public internet, you are already paying the "background radiation tax" in log noise and firewall rules. HoneyPeer lets you **convert that noise into shared immunity**. **[Get free Alpha access →](https://honeypeer.com/pricing)** Questions? **[support.honeypeer.com](https://support.honeypeer.com)** · **[join.honeypeer.com](https://join.honeypeer.com)** · **[docs.honeypeer.com](https://docs.honeypeer.com)** --- *HoneyPeer Alpha is pre-release software. Features, pricing, and terms may change before general availability. Honeypots intentionally attract malicious traffic — deploy only on infrastructure you control and isolate appropriately.*