Boltarr is a self-hosted network dashboard for mapping, monitoring, and

analyzing your local network — all on your own hardware, no cloud.

GitHub: https://github.com/brq-ae/boltarr

Docker Hub: https://hub.docker.com/r/brqae/boltarr

Wiki/Docs: https://github.com/brq-ae/boltarr/wiki

WHAT IT DOES

- Discovery & scanning: nmap-powered host/port discovery with reusable scan

profiles and a scheduler (interval or weekdays, with active-hours windows).

- Change tracking & alerts: records what changed between scans (new hosts,

opened/closed ports, MAC/hostname changes) with ntfy push notifications and

a daily digest.

- Uptime monitoring: host liveness (online/offline) plus per-service up/down

checks with alerts, grace periods, quiet hours, and uptime history — with

host<->service correlation so one down box doesn't flood you with alerts.

- Topology: interactive graph with VLAN coloring and connection types

(wired/wifi/fiber/DAC/virtual).

- Device & service management: add/edit inline, merge multi-homed devices,

DHCP-aware static/dynamic classification, notes, and SSH key management.

- Public status page: push a sanitized status summary (public names + up/down,

never internal IPs) to a separate, internet-facing page — with public/private

gating, branding, announcements, a maintenance calendar, an incident

timeline, and automated-maintenance windows.

- AI assistant (optional): live-network-aware and streaming — ask what's down,

what changed, or what's been flapping. Works with Ollama (local), any

OpenAI-compatible API, or Anthropic. Fully optional; Boltarr works without it.

INSTALL

- Unraid: Community Applications — search "boltarr" (or add the template URL

from the GitHub readme). The template uses Network Type: Host with

NET_RAW/NET_ADMIN caps so nmap can scan your LAN — leave that as-is.

- Docker: a single docker-compose.yml (host networking by default). See the

readme/wiki.

Your data stays local (a small SQLite database + config file). Open source

under Apache-2.0.

Post here for support, bug reports, or feature requests.

Edited August 6Aug 6 by BinAdhed

CHANGELOG & UPDATES (v0.7.0 – v0.17.0)

It's been a big few months since the v0.6.2 port change. Here's everything

that's new, grouped by area, with a full version list at the bottom.

── MONITORING & UPTIME ──────────────────────────────

- Service uptime monitor: per-service up/down checks (HTTP or TCP) with ntfy

alerts, a grace period, quiet hours, and uptime history.

- Host liveness: a light ping sweep marks every host online/offline, tracks

per-host uptime, and correlates service outages to a down host (so one

offline server doesn't fire a flood of service alerts).

- Configurable probe interval (15–600s) and a per-service "reset uptime".

── SCANNING ─────────────────────────────────────────

- Configurable scan engine + reusable scan profiles (Quick/Standard/Full +

custom).

- Scan scheduler with active-hours windows, plus live probe progress.

- Change tracking & alerts: records what each scan finds different (new hosts,

opened/closed ports, MAC/hostname changes) with per-scan summaries and a

daily digest; alert scoping by static/dynamic and per-host opt-outs.

- Timezone-aware scheduling, quiet hours, and digests.

── PUBLIC STATUS PAGE (new, separate app) ───────────

- Push a sanitized status summary to a separate, internet-facing page — public

names and up/down only, never internal IPs. Ships as a prebuilt image.

- Public/private section gating with a hardened admin login, announcements,

and a recurring maintenance calendar.

- Full branding (name, logo, colors), configurable uptime windows

(1h / 24h / 7d / 30d), and a per-minute "1 hour" tick bar.

- Incident timeline (down -> back -> duration), and automated-maintenance

windows that relabel expected dips (e.g. nightly backups) as amber

"Scheduled maintenance" instead of red — the downtime still counts, only the

cause is labelled. Plus a "clear past incidents" control.

── AI ASSISTANT ─────────────────────────────────────

- The assistant is now live-network-aware: it knows what's down right now,

recent changes, and outage/flapping patterns — and it streams responses.

- Still optional and provider-agnostic: Ollama (local), any OpenAI-compatible

API, or Anthropic.

── UI / QUALITY OF LIFE ─────────────────────────────

- Inline editing + column controls in the Hosts and Services tables.

- Topology zoom/ESC improvements; fixed-height, drag-resizable detail panels.

- Factory reset, bulk delete, and an accordion settings panel.

- Editable host IP; restore now migrates backups from older versions.

── UPGRADE NOTES ────────────────────────────────────

- Deployment is now a single docker-compose.yml, and host networking is the

default so nmap can actually scan your LAN. On Unraid the template already

sets Network Type: Host with NET_RAW/NET_ADMIN caps — leave that as-is.

- The public status page runs as its own container (default port 12102) on a

SEPARATE host from Boltarr — keep Boltarr off any internet-facing box.

── FULL VERSION LIST ────────────────────────────────

v0.7.0 service uptime monitor + public status page

v0.8.0 configurable scan engine + reusable scan profiles

v0.9.0 scan scheduler, change tracking & alerts, host liveness, timezone

v0.10.0 scan active windows, alert scoping, live probe progress

v0.11.0 host monitoring (uptime, correlation, status page)

v0.11.2 consolidated to a single docker-compose.yml

v0.11.3 factory reset, bulk delete, accordion settings

v0.12.0 inline editing in the Hosts table

v0.13.0 inline editing in the Services table

v0.14.0 reset uptime tracking

v0.15.0 90-day uptime history on the status page

v0.16.x incident timeline, configurable uptime windows, configurable probe

interval + 1h tick, automated-maintenance windows, clear incidents

v0.17.0 AI assistant is now live-network-aware, with streaming

Docs & full guide: https://github.com/brq-ae/boltarr/wiki

As always, post here for support, bug reports, or feature requests.

Edited August 6Aug 6 by BinAdhed

Is there a quick setup guide that tells step by step what to do and how to get the best use out of it? I set up a subnet and told AI to run an analysis of it (ongoing) but am not really sure I'm getting full use out of it.

On 6/20/2026 at 7:31 AM, dclive said:

Is there a quick setup guide that tells step by step what to do and how to get the best use out of it? I set up a subnet and told AI to run an analysis of it (ongoing) but am not really sure I'm getting full use out of it.

I still did not get to the step by step guide, but it should be easy to get around, what you need to do is first have the app scan the network, it will use nmap built in, once that is done, you can have the Ai (that you connect) analyze your setup.
the point of Ai is after filling up your network, hosts, services etc, the Ai can give you better insight.