Self-hosted · Data flows

What your install sends to nano. And what it never does.

nano runs on your servers and keeps your logs there. This page lists every connection a connected install makes, what each one carries, and how to turn it off.

Buy nano
Talk to sales
The picture

Your logs stay on your side of the line.

Logs are collected, parsed, stored, searched and analyzed on infrastructure you run. What crosses to nano is licensing and one daily number; what comes back is signed content you can mirror or switch off.

Your environment

Your data center, your cloud account, or an air-gapped network

01 · Sources
  • firewalls
  • Windows events
  • identity
  • cloud audit
  • EDR
  • servers
02 · nano
  • Collect and parse
  • Detections and alerts
  • Search and cases
03 · Your storage
  • Disks you own
  • Optional object storage tier
  • Retention you set
Optional · AI pivt AI calls your model endpoint directly: any OpenAI-compatible API (including self-hosted models) or the Anthropic Messages API. Off by default; model calls never go through nano.
nano, the vendor

Never receives your logs.

Connected
  • Daily license check-in: deployment ID, install ID, version
  • Hourly content and enrichment-feed updates (no data sent; can be turned off)
  • One number per day: total GB ingested
  • Downloads: log-source catalog, detection rules and SOAR content, enrichment feeds, releases
Air-gapped
  • No connection to nano at all
  • Offline license file, renewed by hand
fig 01 · What your install exchanges with nano. Log data is never sent.
Sent to nano

Two flows carry anything about you.

Both go to api.nano.rs over HTTPS, opened by your install. nano never connects into your network.

FlowDestinationWhat is sentWhyWhenTurn it off
License check-in GET api.nano.rs/api/license/status nano (api.nano.rs), outbound HTTPS from your install Deployment ID, the running nano version, a random nonce, your license token, and the install's random install ID (X-Nano-Install-Id header). Like any HTTPS request, it also reveals your egress IP address, which nano records with the install. Confirms the license is valid and returns its state, signed by nano. Also returns the newest release, for the in-product "update available" notice and new-version emails. Once a day, and when the API starts. Every 5 minutes until the first success. Use an offline license file instead of the online license values, or air-gapped mode. An online install that can't reach nano keeps its licensed features for 14 days after its last successful check-in, then they lock until it checks in again (see the FAQ below). docs · Network access ↗docs · Installs per license ↗
Usage report POST api.nano.rs/api/license/usage nano (api.nano.rs) One number per completed UTC day: total GB ingested (10^9 bytes, rounded to 0.1), up to 14 days per report, with the deployment ID and install ID. nano's API accepts nothing else about ingested data: no log content, source names, hosts, IP addresses, users, event types, per-source breakdowns or event rates. Capacity and billing: compares usage with your plan (Self-run Addendum, Section 6.4). It never blocks or slows ingest. Once a day, after a successful check-in. No separate switch while you use an online license. With an offline license or air-gapped mode nothing is reported: nano keeps the daily record locally and you provide it at renewal. docs · Network access ↗Self-run Addendum ↗
Connected install with an online license, default settings.
Downloaded from nano

Signed content in, nothing about your data out.

These requests fetch content. Apart from a license credential where noted, they carry nothing about your install beyond what any HTTPS request reveals: your egress IP address and the time.

FlowDestinationWhat is sentWhyWhenTurn it off
Content channels GET api.nano.rs/api/content/v1/… nano (api.nano.rs), then a short-lived download link on Cloudflare R2 (*.r2.cloudflarestorage.com) Open channels (log-source parsers, detection rules, playbooks, enrichment and integration catalogs, the model catalog and MITRE ATT&CK): nothing beyond the request itself. The SOAR catalog (a licensed channel, included with every license): your license token and deployment ID, or your offline license file. Delivers signed content. The install verifies each bundle's signature before using it. A new catalog entry is offered to you; nothing is onboarded on your behalf, and every install starts with zero log sources. Every hour (minimum 5 minutes). CONTENT_CHANNELS_AUTO_UPDATE=false stops all channels; each channel can also be opted out on its own. CONTENT_CHANNELS_URL points at your own mirror. Without updates, nano keeps running on the content it has. docs · Network access ↗
Enrichment feeds GET api.nano.rs/api/feeds/v1/… nano (api.nano.rs), then Cloudflare R2. Not the upstream providers. Nothing beyond the request itself for an install with an online license. An install with an offline license file sends that file to fetch the licensed feed. One signed bundle with IPinfo Lite (IP country, continent and network owner), ThreatFox indicators (abuse.ch) and the Tor Project's exit-node list. Events are enriched locally, at ingest. No event data goes out. Every hour (minimum 5 minutes). On by default. PLATFORM_FEEDS_AUTO_UPDATE=false stops the download (there is no per-feed switch); PLATFORM_FEEDS_URL points at your own mirror. The last bundle keeps enriching until it expires. docs · Platform feeds ↗
Releases and images install and upgrade only Bundle links on storage.googleapis.com; nano images from us-central1-docker.pkg.dev and ghcr.io; ClickHouse, PostgreSQL, Valkey and nginx from Docker Hub; etcd from quay.io (multi-server) Your license token and deployment ID to api.nano.rs, to get download links and a registry credential. Image pulls send only registry credentials. Installs and upgrades nano from signed, digest-verified images. When you install or upgrade. On Kubernetes, a registry-credential refresh job also calls api.nano.rs every 20 minutes when pulling from nano's registry with an online license. Load the signed image bundle into your own registry. On Kubernetes, set registryCredentialRefresh.enabled: false or use your image mirror. docs · Network access ↗docs · Kubernetes ↗
Content and feed updates are on by default. Every bundle is signed and verified before it is used.
Off by default

Everything else waits for you to switch it on.

None of these run on a fresh install. Each connects only to the endpoint you configure, and none of them goes through nano.

FeatureDestinationWhat is sentPurposeDefault
pivt AI The model endpoint you configure Prompts and the investigation context pivt sends AI features, if you turn them on Off until you configure a provider. There is no nano-managed default: point it at any OpenAI-compatible endpoint (including self-hosted vLLM, Ollama or LM Studio) or the Claude (Anthropic Messages) API.
Threat-intel lookups The provider you install (for example URLhaus, VirusTotal, GreyNoise, AbuseIPDB) The artifact being looked up (an IP, domain, URL or hash) with your API key On-demand lookups during investigations Off until an admin adds a provider and its credential. These go straight to the provider, not through nano.
Everything you connect Your identity provider, SaaS APIs, notification channels, S3-compatible storage, GitHub for detection as code What that integration needs SSO, pull feeds, alerts, storage tiering and backups Only when you configure them, to the endpoints you choose.
Connected, offline license, air-gapped

Three ways to run, three different footprints.

Pick the mode that matches your network. All three are included with every license; air-gapped trades the automatic updates for manual imports.

Connected (online license)Offline license fileAir-gapped (AIRGAP_MODE=true)
License check-in Daily None: checked locally against the signed file None: checked locally against the signed file
Daily GB number Sent after each check-in Kept locally; you provide it at renewal Kept locally; you provide it at renewal
Content and feed updates Hourly from nano, unless turned off Hourly from nano (sending the license file), unless turned off None. Signed offline bundles you import
Releases From the registry, or the signed image bundle From the registry, or the signed image bundle Signed image bundle in your own registry
AI Your provider, if configured Your provider, if configured An endpoint inside your boundary, if configured
Air-gapped mode stops nano's own egress. It doesn't block destinations you configure yourself, such as webhooks, SMTP or an external AI provider.
Mirror it

Keep updates, lose the direct line.

Set CONTENT_CHANNELS_URL and PLATFORM_FEEDS_URL to an HTTPS mirror you control. Bundles are signed by nano and verified by the install, so the mirror can't alter them. Network access

Proxy it

One egress point you can watch.

License check-in and content updates go through your corporate HTTP(S) proxy when you set one, so every call to nano shows up in your own proxy logs. Corporate proxy

Read it

Check the code, not just this page.

nano's engine is open source under AGPL-3.0, so you can read how the open edition behaves on the network. Open source

FAQ

Questions we get asked.

Does nano send my logs home?

No. A connected install sends nano a daily license check-in (deployment ID, install ID, version, license token and a nonce) and one number per completed UTC day: the total GB ingested. nano's API accepts nothing else about ingested data. Log content, hosts, users, IP addresses from your events, source names and event rates are never sent.

What does a connected install download from nano?

Every hour by default it checks api.nano.rs for signed content (log-source parsers, detection rules, playbooks, catalogs, MITRE ATT&CK and the SOAR catalog) and for one signed enrichment bundle with IPinfo Lite, ThreatFox and the Tor exit list. Downloads come from Cloudflare R2 links nano hands out. Each bundle's signature is verified before use. You can stop both, or point them at your own mirror.

Can I turn off the default enrichments?

Yes. Set PLATFORM_FEEDS_AUTO_UPDATE=false to stop downloading the IPinfo Lite, ThreatFox and Tor exit-node bundle; the last bundle keeps enriching until it expires. There is no per-feed switch. Enrichment itself happens on your servers: events are looked up in local ClickHouse dictionaries and no event data goes out. URLhaus and similar services are optional lookups that only run if you add them with your own key.

Does nano use its own AI model or gateway?

No. AI is off until you configure a provider, and there is no nano-managed default. pivt calls the endpoint you configure directly: any OpenAI-compatible API, including a self-hosted model on vLLM, Ollama or LM Studio, or the Claude (Anthropic Messages) API. Prompts go to that provider, not through nano.

Can I run nano with no connection to nano at all?

Yes, in air-gapped mode, which is included with every license. With AIRGAP_MODE=true there is no license check-in, no usage report and no content or feed polling. The license is a signed offline file you renew by hand, and content, enrichment data and releases arrive as signed bundles you carry across the gap and import by hand. Air-gapped mode doesn't block destinations you configure yourself, such as webhooks or an external AI provider, so keep those inside your boundary.

What does an offline license change?

An offline license replaces the daily check-in and the usage report: validity is checked locally against the signed file, and nano keeps the daily GB record for you to provide at renewal. On its own it does not stop content and enrichment updates, which then present the license file to fetch licensed content. To stop those too, turn off the two auto-update settings or use air-gapped mode.

Why does nano need my daily ingest volume?

The license is priced per GB/day of ingest. The daily total lets nano compare usage with your plan under Section 6.4 of the Self-run Addendum. It is one number per day, tied to your deployment ID and a random install ID, so the two installs a license covers can be added up. Ingest is never blocked or slowed because of it.

Is there any telemetry, analytics or crash reporting?

No analytics, telemetry or crash-reporting service is built into the nano server or web app. Health metrics are exported only if you set your own OTLP endpoint, and the app forwards its own logs only to a destination you configure.

What happens if my install can't reach nano?

With an online license, an install that can't reach nano keeps its last verified license for 14 days, then its licensed features lock until it checks in again. While locked, pushed ingest (HTTP, HEC, OTLP), search over stored data and scheduled detection rules keep running and local admins can still sign in; SSO sign-in, pull collectors, cases, SOAR and AI stop. No data is deleted or altered. An offline license file is checked locally, so it removes the check-in dependency. Content and enrichment stay at their last versions. If your network can't reach api.nano.rs, use an offline license or air-gapped mode instead.

Sources

Read the details.

Start with nano.

From US$149/month at 5 GB/day, US$279/month at 10 GB/day (list price, billed monthly; 10% off billed annually). Every feature in one license, and one license covers two installs.

Buy nano
Talk to sales