Setup Options

Choose the Rails Yield deployment model that fits your stack — Fully Managed, Third Party Hosted, or Intelligence API — and the permissions and trade-offs of each.

Rails Yield supports three deployment models. They differ in who runs the Lightning node, who runs the automation, and how much access Amboss has to your infrastructure. Pick by your operations posture, not by feature - all three earn the same yield.

Decision tree

Comparison

Option1. Fully Managed2. Third Party Hosted3. Intelligence API
Who runs the nodeAmboss (Voltage-hosted LND/Litd)YouYou
Who runs automationAmbossAmbossYou
Amboss → your node connectionDirect (Amboss-provisioned)Inbound from Amboss to your nodeNone
Integration liftLowestMedium (LND/Litd deployment + macaroon submission)Highest (you build automation)
Best forTreasuries and teams who want hands-off yieldTeams with existing Lightning infrastructureTeams with sophisticated in-house automation
Key benefit✅ Completely hands-off✅ After node setup, almost hands-off✅ No external connectivity to your node
Key trade-off❌ Opinionated stack❌ DevOps requirements❌ Highest integration effort

All three models are self-custodial. Channel keys stay on your node (or the Voltage-hosted node provisioned in your name).

Detailed pages

Node permissions for Rails

Fully Managed and Third Party Hosted setups both grant Amboss a scoped macaroon. The macaroon avoids spend-related permissions so the team retains full control over funds.

These permissions are intentionally scoped to limit operational access while enabling Amboss to perform essential functions on behalf of the user.

Required?Permission
General view-only permissions
Open and close channel permissions
Set routing policy permissions
Create invoice permissions
Create on-chain address permissions
Pay to invoice or on-chain spend

The Intelligence API setup does not connect to your node and therefore requires no macaroon.

Onboarding

For Fully Managed, steps 1-3 are the dashboard "Create Earn Node" flow; for Third Party Hosted, step 3 is the node.connect mutation; for Intelligence API, you stop at step 1 and start querying.

Offboarding

See Withdraw for the API call and Node Lifecycle for the deactivation states.

Next steps