Namera is a programmable wallet infrastructure that lets agents execute onchain transactions within rules you define. Instead of giving full wallet access, you define permissions upfront and let agents operate within those boundaries.

How do agents access wallets?

Agents don't access your wallet directly. They use session keys — secondary keys with limited permissions — that you issue and control.

Are funds safe if agents can execute transactions?

Yes. Agents can only act within the limits you define — like which contracts they can call, how much they can spend, and how often. They never have unrestricted access to your funds.

What are onchain policies?

Policies are rules attached to a session key that define what an agent is allowed to do. They're enforced directly onchain, so they can't be bypassed or modified by the agent.

What happens if an agent is compromised?

Its access is already restricted by policies. You can revoke the session key instantly without affecting your main wallet or other agents.

Do I need to run infrastructure?

No. Namera is local-first and can run directly in your environment via CLI or MCP server. You don't need to manage backend infrastructure to get started.

How do I integrate Namera?

You can use the SDK for full integration, the CLI for local workflows, or connect directly to agents through the MCP server. It's designed to plug into existing developer tools and agent environments.

What is the MCP server?

The MCP server lets AI agents connect to Namera and execute actions safely. It acts as the bridge between your agent and the wallet, enforcing permissions during execution.

Can I use multiple agents with different permissions?

Yes. You can create multiple session keys, each with its own policies, and assign them to different agents or roles.

Which chains does Namera support?

Namera supports multi-chain execution through a unified interface, so you can operate across networks without managing separate wallets.

Introduction

Namera CLI gives agents controlled access to smart accounts with session keys and scoped permissions, running fully local and agent-first across chains.

Namera CLI gives agents controlled access to smart accounts with session keys and scoped permissions, running fully local and agent-first across chains. It is designed to let autonomous systems safely interact with wallets without exposing private keys or relying on centralized infrastructure.

What Namera CLI provides

Session Key based Access: Delegate wallet capabilities to agents using session keys instead of exposing primary private keys.
Scoped Permissions: Define exactly what an agent can do—limit contracts, methods, spending, and time windows.
Local-First Execution: Everything runs on-device with no required cloud dependency, ensuring privacy and reliability.
Agent-Controlled Workflows: Enable agents to execute transactions, automate flows, and interact with smart contracts autonomously.
Multichain Support: Operate across multiple chains seamlessly with a unified interface for signing and execution.
Programmable Permissions: Build complex permission logic with different types of policies.

Agent-first by design

Namera CLI is built around a simple idea: agents should be first-class users of wallets, not an afterthought. This means:

Every command has a JSON schema
Every command can be driven via --params
Prompts are the default, but never required
Outputs are consistent and machine-friendly

How it works

Each command supports two execution modes:

Interactive (default): missing inputs are prompted for
Params mode: provide JSON inputs using --params

Agents can discover the exact input shape using:

namera schema <command>

Example:

namera schema session-key.create

Global flags

These flags are available on every command.

--params: Raw JSON parameters for agent-driven execution
--output, -o: pretty (default), json, or ndjson
--quiet, -q: Suppress output entirely

Params mode

Use --params to bypass prompts and send a full JSON payload. This is the preferred mode for agents.

Example:

namera smart-account create --params '{"alias":"my-smart","ownerAlias":"my-owner","ownerPassword":"my-password","index":0}'

If --params is provided, the CLI ignores any interactive prompts and validates the payload against the schema.

Output formats

Most commands return structured data. Control the output format with --output:

pretty: Human-friendly output (default)
json: Single JSON document
ndjson: One JSON object per line

Use --quiet to suppress output if you only need the exit status.

Quick start

Explore command groups: