Partner program

Build agents your clients can actually deploy.

You build the intelligence. SweetHive supplies the boundaries: scoping, audit, group targeting and on-prem runtime: the parts that get agent projects past IT and legal.

The problem you know too well

The agent isn't the blocker. The deployment surface is.

Every agent consultancy has lived this meeting: the demo lands, the client is excited, then someone asks “where does the data go?” and the project dies in security review. The blocker is rarely your agent. It's the deployment surface: flat permissions, no audit trail, no answer to “what can it see, for whom?”

What SweetHive gives your practice

The parts you shouldn't have to build.

The scoping law, for free

Every agent reads its scope ∩ its owner's visibility, checked on every request and enforced in the platform's retrieval layer. You never build permission logic again, and your client's IT can verify it.

Personal by design

Every member runs their own agent with its own scoped token, so no agent reads beyond what its owner can see. External partners, insurers, parents: separated by construction, for mixed-trust audiences.

On-prem runtime

Install the SweetHive Agents Server on the client's own hardware: every member's agent runs there with local models. “Inference never leaves your infrastructure” becomes a sentence you can say in the sales meeting.

MCP-native

The SweetHive Agents Node is an MCP server for Claude Desktop and Claude Code, and can plug MCP servers you already run into a member's agent as tools, always within that agent's scope.

Audit as a deliverable

Every run logged with sources, effective scope and where it was computed. Your engagement report writes itself.

How the program works

Build, deploy, earn.

1 · Build

Build an app with the SweetHive App SDK and register it under Developer → Apps. It runs inside a context with a short-lived token scoped to that context and to the person using it.

2 · Deploy

Set up the client's hive (contexts, groups, invitations), add your app to the contexts that need it once it is approved for the store, and install the Agents Server on their hardware when inference must stay on-prem.

3 · Operate & earn

Help the client's members set up their personal agents and workflows under the client's policies, and bill your services. Apps can carry their own pricing plans, which clients subscribe to from SweetHive billing.

Program terms, tiers and revenue share: to be defined.

Who we're looking for

Partners who ship into serious environments.

Our first proof vertical is art & heritage, via the AerariumChain HUB, and the same two-tier pattern extends to every regulated, high-trust sector.

  • ⬡AI agent consultancies and automation agencies.
  • ⬡System integrators serving regulated and high-trust sectors.
  • ⬡Vertical builders in heritage, education and professional services.

Ship the agent. Skip the security fight.

Build on the boundaries your clients' IT and legal teams already ask for.

Apply to partner