What can my AI agent do with Bron?

Bron gives you an exceptionally powerful set of tools for working with your agent. You decide exactly what your agent can do: our agent tooling lets you set the level of access and the capabilities your AI agent has within your wallet. These tools allow your agent to perform any action you could perform yourself, yet every transaction still requires your explicit approval before it goes through. You don’t need a separate device for this—all approvals can be handled through our desktop or mobile applications. If you’d prefer your agent to run fully autonomously, you can set up an MPC hot signer. Keep in mind that, in this configuration, your agent will be able to execute transactions without your explicit approval.

How do I make sure my agent only does what I’ve approved?

With Bron’s MPC architecture, you can connect your AI agent to your wallet and assign it the role of transaction operator. The agent can read all wallet data and submit transactions for signing, but it cannot execute transactions on its own. You approve or cancel every transaction from your laptop or phone, so you always retain full control over the agent and explicitly authorize each action it takes. No additional hardware is required—every approval is handled effortlessly through the app on your phone or laptop.

Does Bron hold or control my agent’s on-chain assets?

Bron is self-custodial software, so it never stores any of your funds or on-chain assets. Bron gives your agent the tools to interact with the blockchain, and gives you control over the agent at whatever level of granularity you require. No assets ever pass through Bron.

Which AI frameworks does Bron support?

Bron ships a native MCP server compatible with Claude, ChatGPT, and other MCP-enabled assistants. For scripted workflows, a full-featured CLI is available. For custom integrations, use the REST API at api.bron.org.

How do I connect my AI agent to Bron?

Today this is a developer setup: you install Bron’s bron CLI on your computer and drive it from an LLM agent — Claude Code, Cursor, or custom tooling — directly, or through its built-in MCP server. State-changing actions like approving or signing always require your confirmation, and you can set guardrails such as spend limits and dry-runs before anything goes through. A simpler, in-app way to connect your agent — with a fuller step-by-step guide — is coming soon. For now, full instructions plus the CLI, REST API, and SDKs are at developer.bron.org/sdk/cli/agents.

The Wallet Agents Deserve

Bron Wallet at the centre of a network of AI agents

You Hold the Veto

AI is fast, but occasionally wrong. The Bron policy engine gives you address whitelists, spend limits, per-transaction approval and more. Stay in control of your agent from your phone or laptop. No additional hardware to carry, no seed phrase to lose.

No Tech Knowledge Needed

Native MCP server for Claude, ChatGPT and others. Fully-featured CLI for scripted flows and a clean API for everything else. Whatever your agent runs on, Bron speaks the language. Simply run a ready-made prompt and everything works.

Get started

Keyboard keys carrying the logos of AI assistants

Prompts That Save Hours

Scan every chain for incoming payments, claim and re-stake rewards, trade on DEXs, run end-of-month payroll from a team sheet, or reconcile the books for your accountant. Create hundreds of flows while staying in control. Explore Bron's agent tools — MCP, CLI and SDKs — at developer.bron.org.

Enterprise Agent Capabilities at the Price of a Lunch

Full policy engine, MCP, CLI, API, access to swaps, DEX connectivity and even an inheritance plan for your agent. All included in the Basic tier, no strings attached.

FAQ

: Bron gives you an exceptionally powerful set of tools for working with your agent. You decide exactly what your agent can do: our agent tooling lets you set the level of access and the capabilities your AI agent has within your wallet. These tools allow your agent to perform any action you could perform yourself, yet every transaction still requires your explicit approval before it goes through. You don’t need a separate device for this—all approvals can be handled through our desktop or mobile applications. If you’d prefer your agent to run fully autonomously, you can set up an MPC hot signer. Keep in mind that, in this configuration, your agent will be able to execute transactions without your explicit approval.
: With Bron’s MPC architecture, you can connect your AI agent to your wallet and assign it the role of transaction operator. The agent can read all wallet data and submit transactions for signing, but it cannot execute transactions on its own. You approve or cancel every transaction from your laptop or phone, so you always retain full control over the agent and explicitly authorize each action it takes. No additional hardware is required—every approval is handled effortlessly through the app on your phone or laptop.
: Bron is self-custodial software, so it never stores any of your funds or on-chain assets. Bron gives your agent the tools to interact with the blockchain, and gives you control over the agent at whatever level of granularity you require. No assets ever pass through Bron.
: Bron ships a native MCP server compatible with Claude, ChatGPT, and other MCP-enabled assistants. For scripted workflows, a full-featured CLI is available. For custom integrations, use the REST API at api.bron.org.
: Today this is a developer setup: you install Bron’s bron CLI on your computer and drive it from an LLM agent — Claude Code, Cursor, or custom tooling — directly, or through its built-in MCP server. State-changing actions like approving or signing always require your confirmation, and you can set guardrails such as spend limits and dry-runs before anything goes through. A simpler, in-app way to connect your agent — with a fuller step-by-step guide — is coming soon. For now, full instructions plus the CLI, REST API, and SDKs are at developer.bron.org/sdk/cli/agents.

Curious to
delve further?

Bron wiki

Your Access
Starts Now