Graph-first resolution
A stock recursive resolver answers every query it can - which is exactly the problem.
It has no opinion about a name registered six minutes ago, hosted on bulletproof infrastructure, behind a hijacked prefix, or owned by an entity you've explicitly banned. It resolves that just as happily as it resolves github.com, and your agent connects.
Static blocklists chase that gap and always lose - a feed refreshed hourly is still hours behind a fast-flux C2 domain or a newly announced /24 that just hijacked someone else's routes. Whisper's resolver closes the gap a different way: it doesn't consult a list, it asks a live graph - the same one behind RDAP and whisper.agents - a question with real teeth: who owns this, where does it sit, what category is it, is its route valid right now? - before it lets a single query leave for the wider internet.
Two resolvers, same wire protocol
Nothing about the wire format changes. dig, kdig, and every stub resolver on earth still send the same RFC 1035 query and expect the same RFC 1035 response - a 12-byte header, a question section, and (on success) an answer section of resource records. The difference is what happens between the query landing and the answer leaving.
With stock tools, against a normal resolver - it forwards blindly, no matter what's on the other end:
$ dig @1.1.1.1 +noall +comments +answer AAAA some-freshly-registered-domain.example
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 41232
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1
some-freshly-registered-domain.example. 300 IN AAAA 2606:4700::...
1.1.1.1 has no idea the name was registered nine minutes ago, or that its authoritative nameserver sits behind an ASN with a week-old RPKI-invalid announcement. It isn't supposed to know - that's not its job. It just does RFC 1035.
With Whisper, against an agent's resolver - same query, same wire format, but the answer reflects a policy decision made milliseconds earlier:
$ dig @2a04:2a01:0:53::53 +dnssec +noall +comments AAAA some-freshly-registered-domain.example
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 41232
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
Same source address, same query - but this tenant's policy blocks newly-registered as a category, so the name never reaches a real answer. No packet toward that domain's infrastructure ever left the agent's /128. The block is logged the same instant, visible via whisper.agents({op:'logs'}) or the CLI, with the category and rule that fired.
The flow: graph first, real DNS second
Every query an agent sends hits three possible outcomes, evaluated against your policy before resolution proceeds:
| Verdict | What happens | Example trigger |
|---|---|---|
| Allow | Falls through to real resolution - the graph has an opinion and it's positive, or you've explicitly allow-listed the owner/category | An entity you've named on your allow-list |
| Block | Answered on the spot - NXDOMAIN or a fixed sinkhole address, your choice via policy, logged with the rule that matched | Tor exit, bulletproof hosting, newly-registered domain (<48h old) - any category on your block list |
| No opinion | Steps aside immediately - real upstream resolution proceeds at full speed, exactly like a stock resolver | The overwhelming majority of the internet: nothing in your policy or the graph's threat surface says otherwise |
The graph read carries a strict time budget. If it doesn't answer inside that budget, the resolver treats it as no opinion and fails open - a slow or unreachable policy check never becomes a resolution outage. This is deliberate: a control that can black-hole your production traffic when it has a bad day is worse than no control at all. Reliability wins; the query resolves, and the slow check is what gets flagged, not the agent.
Policy dimensions the graph makes possible
A blocklist is a flat set of names. The graph carries structure - ownership graphs, ASN and prefix delegation, RIR geolocation, real-time RPKI validity, and content/category classification - so a policy rule can reach conclusions a list never could. Category is live today: block a whole class of infrastructure - Tor exits, bulletproof hosting, newly-registered domains, known scanning ranges - not by enumerating IPs, but by the graph's classification of what a host is. That's the toggle in the worked example below, and it's one call on the control plane, not a maintenance job.
Geography (geo-fencing resolution to jurisdictions you name), Ownership (denying an entity across every provider it hides behind), and Routing (dropping a destination the instant its route goes RPKI-invalid, per RPKI & routing security) are the shape the graph is built to support and the direction this page is heading - but as of this writing, op:'policy' accepts a geo or routing argument and answers ok:true without persisting either: a follow-up read of the stored policy shows only default, mode, retention, and block/allow. Sending them today is a silent no-op, not an error, which is exactly the kind of gap this project doesn't like to leave open - treat only block/allow-by-category as live until this page says otherwise.
Setting policy: dual paths, same effect
With stock tools there is no equivalent - this is the point. A resolver that only speaks RFC 1035 has no channel for "block this category" short of you maintaining and pushing an RPZ zone or a blocklist file yourself, by hand, forever behind the threat.
With Whisper, one call, either as the raw control verb or the CLI wrapper:
# raw control plane - the whisper.agents verb, over the graph gateway, with your key
curl -s https://graph.whisper.online/api/query \
-H "X-API-Key: whisper_live_..." -H 'content-type: application/json' \
-d '{"query":"CALL whisper.agents({op:\"policy\", block:[\"tor-exit\",\"newly-registered\"]})"}'
# the CLI, same effect
whisper policy --block tor-exit --block newly-registered
Live, read back with no arguments:
$ whisper policy
KEY VALUE
default allow
mode hybrid
retention 90
block tor-exit
block newly-registered
...
From that point forward, every query from my-agent's address is measured against those rules before it resolves - verify it live:
dig @{agents-53-resolver} +short AAAA {a-known-suspicious-test-name}
# -> NXDOMAIN, and the block shows up in `whisper.agents({op:'logs'})` immediately
Two access paths, one policy engine
The graph-first check applies identically whether an agent reaches Whisper over DNS-over-HTTPS (RFC 8484, :443, API-key header - good for HTTP-only egress environments) or a dedicated per-tenant IPv6 /128 resolver on standard :53, where the source address of the query is the authentication - no key on the wire at all:
# DoH - keyed, wire-format DNS-over-HTTP
curl -s -H "X-API-Key: whisper_live_..." -H 'content-type: application/dns-message' \
--data-binary @query.bin https://doh.whisper.online/dns-query
# per-tenant :53 - keyless, the source /128 is the identity
kdig @2a04:2a01:0:53::53 +dnssec example.com AAAA
Both paths run the same three-way verdict, the same fail-open timeout, and the same per-query log line - pick the transport that fits your agent's egress path, not the policy.
Your resolver is a pair
Your dedicated /128 has a twin: the same address with bit 64 flipped, bound on the other server of the lockstep pair. Each server binds one member, so when either one restarts, the other member keeps answering. op:resolver returns both, primary first, and local_address names the member bound on the server that took the call. Two calls a moment apart landed on different servers:
for i in 1 2; do curl -s https://graph.whisper.online/api/query -H "X-API-Key: whisper_live_..." \
--data-urlencode "q=CALL whisper.agents({op:'resolver'})" | jq -c '.rows[0] | {ok, result}'; done
{"ok":true,"result":{"columns":["address","state","addresses","local_address"],"rows":[["2a04:2a01:0:53:ca61:46b0:e0c7:ae90","active",["2a04:2a01:0:53:ca61:46b0:e0c7:ae90","2a04:2a01:0:53:4a61:46b0:e0c7:ae90"],"2a04:2a01:0:53:4a61:46b0:e0c7:ae90"]],"sensitive":[]}}
{"ok":true,"result":{"columns":["address","state","addresses","local_address"],"rows":[["2a04:2a01:0:53:ca61:46b0:e0c7:ae90","active",["2a04:2a01:0:53:ca61:46b0:e0c7:ae90","2a04:2a01:0:53:4a61:46b0:e0c7:ae90"],"2a04:2a01:0:53:ca61:46b0:e0c7:ae90"]],"sensitive":[]}}
Both members answer, from anywhere, over plain DNS:
$ dig +tcp +short @2a04:2a01:0:53:ca61:46b0:e0c7:ae90 example.com A
172.66.147.243
104.20.23.154
$ dig +tcp +short @2a04:2a01:0:53:4a61:46b0:e0c7:ae90 example.com A
172.66.147.243
104.20.23.154
An outside probe queried every tenant resolver pair once a second across both servers' restarts on 2026-11-10. Each server's own addresses went dark for 73 and 70 seconds; no pair went silent for a single second.
The released CLI (0.264.2) writes both. A real capture in a Debian 12 container, where the profile is plain /etc/resolv.conf, forcing the dedicated :53 profile with --resolver:
# whisper resolver --resolver --label w302-pair
whisper: Tier-2 DNS applied - this machine now resolves through Whisper (your dedicated :53 resolver)
verified every lookup goes to 2a04:2a01:0:53:ca61:46b0:e0c7:ae90 or 2a04:2a01:0:53:4a61:46b0:e0c7:ae90
doh url https://doh.whisper.online/{token}/dns-query
resolver 2a04:2a01:0:53:ca61:46b0:e0c7:ae90, 2a04:2a01:0:53:4a61:46b0:e0c7:ae90 (in order; the second is the failover)
note plain :53 is unencrypted on-path; the DoH profile (default mode) is the encrypted alternative
note the dedicated resolver answers on IPv6; this host needs a global IPv6 route
note two resolver addresses are configured: lookups go to the first, and move to the second if the first stops answering
revert whisper resolver --off
# cat /etc/resolv.conf
nameserver 2a04:2a01:0:53:ca61:46b0:e0c7:ae90
nameserver 2a04:2a01:0:53:4a61:46b0:e0c7:ae90
options timeout:1 attempts:2
search botboss.app
# whisper resolver --off
whisper: resolver profile reverted - this machine's DNS is back to what it was
options timeout:1 attempts:2 keeps the order: the stub asks the primary first and moves to the twin after a second without an answer. There is no rotate, so the load stays on the primary while it is up. whisper resolver --print shows the same pair without changing anything, and --off revoked the resolve-only token the apply minted.
Bare hostnames on your own machines
whisper resolver points the machine it runs on at your dedicated resolver and installs your namespace as a real search domain, so a device you named w190-958-db01 answers to exactly that, with no suffix to type. whisper resolver --off puts the operating system's DNS configuration back the way it was and revokes the resolve-only token the apply minted. Nothing below was set up by hand: the suffix comes from the account itself.
These are real captures from the released CLI (0.255.0) on a stock Ubuntu 24.04 host and on Windows Server 2022, against an account whose devices live under botboss.app. w190-958-db01 is another device on the same account.
Linux, systemd-resolved. Before, the bare name goes nowhere:
$ resolvectl query w190-958-db01
w190-958-db01: resolve call failed: No appropriate name servers or networks for name found
$ sudo whisper resolver --label w190-958-linux
whisper: Tier-2 DNS applied - this machine now resolves through Whisper (your dedicated :53 resolver)
doh url https://doh.whisper.online/{token}/dns-query
resolver 2a04:2a01:0:53:ca61:46b0:e0c7:ae90
note Linux default: your dedicated per-tenant resolver on :53 - no forwarder needed (force the encrypted DoH forwarder profile with --doh)
note plain :53 is unencrypted on-path; the DoH profile (default mode) is the encrypted alternative
note the dedicated resolver answers on IPv6; this host needs a global IPv6 route
revert whisper resolver --off
$ cat /etc/systemd/resolved.conf.d/whisper-resolver.conf
# Whisper resolver - generated by `whisper resolver`; revert with `whisper resolver --off`.
[Resolve]
DNS=2a04:2a01:0:53:ca61:46b0:e0c7:ae90
Domains=~. botboss.app
$ resolvectl status
Global
Protocols: -LLMNR -mDNS -DNSOverTLS DNSSEC=no/unsupported
resolv.conf mode: stub
Current DNS Server: 2a04:2a01:0:53:ca61:46b0:e0c7:ae90
DNS Servers: 2a04:2a01:0:53:ca61:46b0:e0c7:ae90
DNS Domain: botboss.app ~.
...
$ resolvectl query w190-958-db01
w190-958-db01: 2a04:2a01:ba1b:2471:5afd:5baf:d9d0:4382 -- link: enp1s0
(a5afd5bafd9d04382.botboss.app)
$ host -t AAAA w190-958-db01
w190-958-db01.botboss.app is an alias for a5afd5bafd9d04382.botboss.app.
a5afd5bafd9d04382.botboss.app has IPv6 address 2a04:2a01:ba1b:2471:5afd:5baf:d9d0:4382
$ ping -6 -c1 -W2 w190-958-db01
PING w190-958-db01 (2a04:2a01:ba1b:2471:5afd:5baf:d9d0:4382) 56 data bytes
$ resolvectl query w190-958-not-a-member
w190-958-not-a-member: Name 'w190-958-not-a-member' not found
Domains= carries both forms on purpose. ~. is a routing domain: every query goes to your resolver and nowhere else. The bare botboss.app is the search domain that turns w190-958-db01 into w190-958-db01.botboss.app. systemd-resolved never sends a single-label name to unicast DNS by itself, so on these hosts the search domain is what makes a bare name work at all. A packet capture during the queries above showed every port-53 packet going to the tenant resolver /128 and to nothing else. On a host without systemd-resolved, the same pair lands in /etc/resolv.conf as nameserver and search.
The revert:
$ sudo whisper resolver --off
whisper: resolver profile reverted - this machine's DNS is back to what it was
$ resolvectl query w190-958-db01
w190-958-db01: resolve call failed: No appropriate name servers or networks for name found
A snapshot of /etc/resolv.conf, every file under /etc/systemd/resolved.conf.d/ with its mode and hash, and resolvectl status, taken before the apply and again after the revert, diffs to nothing.
One caveat on Linux, measured rather than assumed: if another file under resolved.conf.d/ already sets DNS=, systemd-resolved merges the lists and keeps using the server that file named first. The search domain still works, so the bare name resolves, but the queries go to that other server rather than to yours. Check Current DNS Server in resolvectl status after the apply.
Windows, in an elevated PowerShell. The profile is encrypted DoH, steered by one NRPT rule, and your suffix is added to the global suffix search list beside any an administrator has already set:
PS> Resolve-DnsName w190-958-db01 -Type AAAA
w190-958-db01 : The filename, directory name, or volume label syntax is incorrect
PS> whisper resolver --label w190-958-win
whisper: Tier-2 DNS applied - this machine now resolves through Whisper (encrypted DoH)
doh url https://doh.whisper.online/{token}/dns-query
resolver 2a04:2a01:0:53:ca61:46b0:e0c7:ae90
note run in an ELEVATED PowerShell (DNS changes need elevation)
revert whisper resolver --off
PS> Get-DnsClientGlobalSetting | Select-Object -Expand SuffixSearchList
botboss.app
PS> Get-DnsClientNrptRule | Format-List Namespace,NameServers,DnsOverHttps,Comment
Namespace : {.}
NameServers : {78.141.218.153, 45.32.186.184, 2a04:2a00::80}
Comment : whisper-resolver
PS> Resolve-DnsName w190-958-db01 -Type AAAA
Name Type TTL NameHost IPAddress
---- ---- --- -------- ---------
w190-958-db01.botboss.app CNAME 60 a5afd5bafd9d04382.botboss.app
a5afd5bafd9d04382.botboss.app AAAA 60 2a04:2a01:ba1b:2471:5afd:5baf:d9d0:4382
PS> ping -6 -n 1 w190-958-db01
Pinging a5afd5bafd9d04382.botboss.app [2a04:2a01:ba1b:2471:5afd:5baf:d9d0:4382] with 32 bytes of data:
PS> Resolve-DnsName w190-958-not-a-member -Type AAAA
w190-958-not-a-member : DNS name does not exist
PS> whisper resolver --off
whisper: resolver profile reverted - this machine's DNS is back to what it was
--off removes only what the apply added: our NRPT rule (found by its whisper-resolver comment) and our suffix, leaving an administrator's own suffixes in place. The suffix search list, the NRPT table, the DoH server table, each interface's DNS servers and the registry SearchList read the same after the revert as before the apply.
Next
- Egress governance - the same graph-first check applied to outbound connections, not just names
- DNS-over-HTTPS - the keyed
:443transport for HTTP-only environments