FAQ

Frequently asked questions about the Obol Stack

General

What is the Obol Stack?

The Obol Stack is a local Kubernetes-based environment for running AI agents alongside blockchain infrastructure. It uses k3d (Kubernetes in Docker) to provide a lightweight cluster where you can deploy a default AI agent (Hermes), Ethereum nodes, Layer 2 networks, payment-gated services, and other applications.

Is the Obol Stack suitable for production?

It is not yet advised. The Obol Stack works on mainnet, as do DV pods, but it would be wise to only use it with less than a threshold amount of validator nodes, or with only as much crypto assets under management that you would be okay with an Obol Agent losing. For production validator deployments, see the Run a DV docs for the latest mainnet guides.

What operating systems are supported?

  • macOS (Darwin) — Intel and Apple Silicon

  • Linux — amd64 and arm64 architectures

Windows is not currently supported. Windows users may be able to use WSL2, but it is not yet tested.

What are the hardware requirements?

Resource
Minimum
Recommended

CPU

4 cores

8+ cores

RAM

8 GB

16+ GB

Storage

50 GB

500 GB – 2 TB (depends on networks)

Running multiple networks or full Ethereum mainnet nodes significantly increases resource requirements.

Installation

The installer fails with "Docker daemon not running"

Ensure Docker is installed and running:

How do I update the Obol Stack?

Re-run the installer:

The installer will update the CLI binary while preserving your configuration and data. Apply chart updates inside the running cluster with obol upgrade.

How do I uninstall the Obol Stack?

The installer cannot modify /etc/hosts

Manually add the entry:

The Obol Agent

What's the default agent?

Hermes is the default Obol Agent runtime as of v0.9.0. obol stack up provisions a default Hermes instance in the hermes-obol-agent namespace, with its own Ethereum signing wallet and a built-in skill library.

OpenClaw remains supported as an optional alternate runtime — obol agent new --runtime openclaw if you want one.

How do I chat with the agent?

That command passes through to the in-cluster Hermes CLI and gives you an interactive chat TUI. Other useful pass-throughs:

How do I get my agent to message me on Telegram / Discord / Slack?

This walks through Hermes' messaging integrations. Once configured, the agent can ping you when long-running work finishes, when a paid service settles a transaction, or any time it decides it needs your attention.

How do I change the model provider?

Or set up a specific provider:

Custom endpoints work with any OpenAI-compatible server (vLLM, sglang, mlx-lm, etc.).

Where's the agent's Ethereum wallet?

Each agent instance gets a unique Ethereum signing wallet, backed by a remote-signer service. To inspect:

Back it up — losing it means losing the agent's on-chain identity:

Selling services

What is obol sell demo?

obol sell demo deploys a small HTTP service behind an x402 payment gate and prints copy-paste instructions for paying it. It's the canonical "first paid endpoint" experience on the Obol Stack.

Use it to see the full sell → discover → pay → receive loop end-to-end before you wrap your own services with obol sell http or obol sell inference.

Why pay in $OBOL on mainnet?

Buyers paying in $OBOL on Ethereum mainnet sign an EIP-2612 permit off-chain. The Obol-operated x402 facilitator batches that permit with the transfer at settlement time, so buyers never spend ETH on gas and never need to do a one-time approve. They just sign a message, the seller gets paid in OBOL, the facilitator covers the on-chain cost.

USDC and other tokens settle on the rail their issuer supports (EIP-3009 for USDC).

How do I list my service on a public agent registry?

This publishes the agent's wallet + service catalog to the ERC-8004 Identity Registry on the chain you specify. Note that this requires ETH on the registering wallet for gas.

obol sell demo deliberately skips registration by default — run obol sell register later when you want on-chain discovery.

Stack operations

The cluster fails to start

Port conflicts:

The Obol Stack uses ports 80, 443, 8080, and 8443. Check for conflicts:

If ports 80/443 are taken (common on macOS where they require root), edit ~/.config/obol/k3d.yaml and remove the 80:80 and 443:443 entries (keep 8080:80 and 8443:443). Access at http://obol.stack:8080 instead.

Previous cluster not cleaned up:

Can I use my existing kubectl configuration?

The Obol Stack uses an isolated kubeconfig at ~/.config/obol/kubeconfig.yaml. To use it with your standard kubectl:

Or use the bundled passthrough: obol kubectl get nodes.

How do I persist data across cluster restarts?

Data is automatically persisted. obol stack down stops the cluster but keeps data in ~/.local/share/obol/. Only obol stack purge -f deletes persistent data.

Networks

How do I know which deployment ID was generated?

The deployment ID is displayed when you run obol network install. You can also list deployments:

Can I run multiple instances of the same network?

Yes. Each deployment gets a unique namespace:

How long does it take to sync an Ethereum node?

Network
Client
Approximate Time

Hoodi

Reth

2–6 hours

Hoodi

Geth

4–12 hours

Mainnet

Reth

1–3 days

Mainnet

Geth

3–7 days

Sync times depend on hardware, network connection, and chain state.

My network deployment is stuck in Pending state

Common causes: insufficient Docker resources, PVC not bound, image pull errors.

Claude Code plugin

How do I use the Obol skills with Claude Code?

The run-obol-stack skill teaches Claude how to drive the CLI: bring-up, agent setup, debugging, deploying networks, and walking through obol sell demo. Source: github.com/ObolNetwork/skills.

Troubleshooting

Error: "stack not running, use 'obol stack up' first"

Start the cluster:

Error: "deployment not found" or "multiple deployments found"

If you have a single deployment, obol network sync auto-selects it. With multiple deployments, specify which one:

How do I report a bug?

Report issues on GitHub: https://github.com/ObolNetwork/obol-stack/issues

Include:

  • Obol Stack version (obol version)

  • Operating system and architecture

  • Docker version (docker version)

  • Steps to reproduce

  • Relevant logs (obol kubectl logs -n <ns> <pod>)

Need more help?

Head over to our Discord where a member of our team or the community will be happy to assist you.

Last updated

Was this helpful?