ProxyHubb Docs

Provider CLI

Install proxyhubb-cli / provider-node, join with a token, start the tunnel — do not re-register after the wizard.

Providers sell bandwidth by running the provider-node tunnel daemon against proxy.proxyhubb.com:3128. The companion CLI binary is proxyhubb-cli.

Do not confuse provider tunnel :3128 with buyer CONNECT :8090.

  1. Complete the in-app provider setup wizard
  2. Download CLI + provider.toml from CLI Tools
  3. Start the node — do not run join / register again after the wizard (reuse the downloaded config / --node-id)

Downloads

Binaries ship via GitHub Releases and:

http
GET /v1/cli/download?platform=…&arch=…&version=…&package_type=…

The dashboard detects OS/arch and offers the matching asset. Packages are not published to crates.io for day-to-day installs.

npm / pip (same binaries)

Sellers and providers can also install a thin wrapper that downloads provider-cli / provider-node from the API and proxies commands:

bash
# Easiest (Windows / macOS / Linux) — Dashboard → My nodes → Connect → Setup this PC
# Downloads a one-click setup script that writes config, installs the CLI, and starts the node.
 
# npm
npm install -g @proxyhubb/provider@latest
proxyhubb-provider install
proxyhubb-provider --daemon doctor
 
# PyPI
pip install -U proxyhubb-provider
proxyhubb-provider install
proxyhubb-provider --daemon doctor
 
# Cargo (from source — not crates.io)
cargo install --locked --git https://github.com/samuel-1-avson/ProxySocialHub.git --tag provider-v1.0.6 --path services/provider-node

Env overrides: PROXYHUBB_API_URL, PROXYHUBB_CLI_VERSION. Use proxyhubb-provider install to force re-download of the current latest binary, proxyhubb-provider path to print the cached binary, and --daemon to target provider-node instead of provider-cli.

On Windows, if you only download provider.toml into Downloads, recent provider-node builds auto-import it into %APPDATA%\proxyhubb\provider.toml on start/doctor.

There is no separate “seller CLI” — marketplace sellers are providers; automate listings via the SDKs and run bandwidth with this CLI / node.

CLI commands

CommandPurpose
joinEnroll with provider token; write provider.toml
startStart the tunnel daemon
stopStop
statusLocal status
doctorConfig / connectivity checks
dashboardOpen provider dashboard helpers
watchFollow logs / health
installService install helpers
pathsShow config/data paths

Join (fresh enroll)

bash
proxyhubb-cli join \
  --api-url https://api.proxyhubb.com \
  --key <provider_token> \
  --gateway proxy.proxyhubb.com:3128
 
# Reuse a dashboard-created node (skip re-register):
proxyhubb-cli join … --node-id <uuid>

Start / doctor

bash
proxyhubb-cli doctor --config provider.toml
proxyhubb-cli start --config provider.toml
 
# Or run the daemon binary directly:
./provider-node --config provider.toml

Confirm the node goes online under My nodes. Browser “ping” is not a real tunnel — use Connect + CLI start.

Config (provider.toml)

Typical fields (written by join / validated by provider-node):

FieldNotes
control_api_urle.g. https://api.proxyhubb.com
gateway_addr / gateway_addrsTunnel endpoint (host:3128)
jwt_token / api_keyAuth material from join
node_idRegistered node UUID
provider_signing_keyUsage proof signing — keep secret
connection_typeresidential | mobile | datacenter | fiber
use_tlsTLS to gateway when required
metrics_portDefault 9090
max_connections, bandwidth_limit_mbpsCapacity caps

Keep provider.toml private. Never commit signing keys.

Proof signing

If the dashboard shows unsigned usage / “Fix proof”, ensure provider_signing_key matches the key bound at registration. CLI join/register generates it; web-only flows may need the fix banner path.

Common errors

SymptomFix
Node stays offlineRun doctor; check :3128 reachability and token
Accidental second nodeUse wizard config / --node-id; don’t re-join blindly
Unsigned usageAlign provider_signing_key with registered key
Connected to wrong portProviders use 3128, buyers use 8090

Next: Provider nodes — heartbeats, lease, listings.