Developers

Your service, with our governance around it.

Define a tool by hand or import an OpenAPI document. Name, inputs, output, authentication and risk level. It joins the registry and behaves exactly like a built-in capability — permissioned, approvable, rate-limited, audited, revocable.

Read the docs Quickstart

Narrow beats convenient

One endpoint can become one tool or six. Six narrow tools can each be granted to a different agent and refused to the others; one wide tool cannot be reviewed at all.

Your declared risk level is checked

A tool declared low risk that writes to a ledger is caught in review. Declaring risk is your call; being held to it is ours.

Credentials never reach the agent

Your service's authentication is stored as a secret scoped by agent and environment, and injected at call time. It is never in an agent's context, so it cannot be echoed by a model or read from a log.

Schemas are enforced on the way in

Inputs are validated against your declared schema before your service is called. A malformed call fails at the boundary rather than in your code.

Roughly what it looks like

Illustrative and deliberately short. The reference is at docs.agentlabora.com, and pretending a marketing page is documentation wastes an afternoon.

A tool definition

yaml

name: warehouse.check_stock
description: Stock on hand for one product code.
risk: low
inputs:
  product_code: { type: string, pattern: '^[A-Z]{2}-\d{4}$' }
outputs:
  on_hand: { type: integer }
auth: secret://warehouse-api

Where it stops

The platform governs how your tool is called; it cannot make a badly scoped tool safe. A tool that accepts an arbitrary query string has a risk that depends on what the agent asks for, which means it cannot be meaningfully reviewed — and no amount of permissioning around it fixes that.

Developer platform

The rest of it.

QuickstartA deployed agent in under fifteen minutes.SDKsPython and TypeScript, maintained in step.APIAgents, runs, tools and events over HTTP.CLICreate, test, deploy, tail logs, roll back.Agent manifestThe declarative file your agent lives in.SandboxSynthetic data, mock tools, no production reach.EvaluationsTest suites that run in your pipeline.PublishingShip an agent to the marketplace.DocumentationThe reference, the guides and the API explorer.

Give AI a job.

Build an agent, choose what it can access, define what it may do, and deploy it to work safely across your business.

Build an agent Build with the API

No code required · Full SDKs and API · Stop any agent instantly · [email protected]