Run agents free. Govern them when it matters.

Install and run locally

Zero procurement. Point at any OpenAI-compatible endpoint and ship.

pip install kimss-agents

from kimss_agents import Agent, tool

@tool
def multiply(a: float, b: float) -> float:
    """Multiply two numbers."""
    return a * b

agent = Agent(model="gpt-4o-mini", tools=[multiply])
print(agent.run("What is 6 times 7?"))

One line to the Kimss control plane

Same agent code. Production governance when security requires it.

agent = Agent(
    model="custom:your-vaulted-model",
    tools=[multiply],
    gateway="kimss",
    agent_id="ops_bot",
    workspace_key="kimss_...",
)

That sets base_url=https://api.kimss.ai/v1 and sends X-Kimss-Agent-Id. Disable the agent in Kimss and attributed traffic stops (kill switch). See agent kill switch and AI gateway.

What stays free vs what CISOs buy

Open core: the harness and Developer gateway hook stay free; enterprise governance stays paid.

• Free forever (OSS): local agent loop, tools, optional MCP — kimss-agents.
• Free Developer gateway: Invisible Proxy, inventory, audit trail, kill switch, 25k governed requests/month.
• Production+: Authority Boundary, Threat Intercepts UI, Team & Access, PII scrub, argument-policy editing.
• Scale / Enterprise: Entra SSO, SCIM, longer retention, schema isolation, alerts.

Full ladder: /pricing.

Frequently asked questions