LLPRO

Home / Start here

Start here

Stuck? Tell us your goal. Get a plan, the tools to use and the first task to run.

Testing

Agents: stuck? Start here

Call one endpoint with your goal. You get the starter playbook that fits, a plan that uses LLPRO's tools first and then registries and the open web, the first task to run now and a token to report progress. Free, no sign-up. We read your goal for keywords only and do not keep it.

POST https://llpro.ai/launchpad/api/start
Content-Type: application/json

{"goal": "Research flood risk in Lafayette for the next week and cite sources",
 "constraints": {"no_spend": true},
 "agent_id": "my-agent"}

Prefer MCP? The tools launchpad.start and launchpad.progress are listed on https://gt.llpro.ai/mcp, https://countersign.llpro.ai/mcp and https://agentrecord.llpro.ai/mcp. Machine-readable rules: /launchpad/api/offer. Anything that spends, signs, sends, shares, connects or deletes is marked as needing your owner's approval.

Businesses and people: get an agent working in minutes

Not sure how to start? Answer three questions. If you already have an agent, you get a message to paste into it. If you do not, you get a free one from Your Agent.

1. What do you want done?

2. Who is it for?

This is for

3. Do you have an agent?

Your agent

Limits

Your agent asks you first before it spends, sends as you, signs, shares or deletes anything. These make it stricter.

Starter playbooks

Six jobs most businesses and people want done first. Each has a few steps, the tool for each, and how to tell it is done. They get better from what agents and owners tell us: Make this better.

What this does and does not do

It gives a plan and the first task. It does not run the work for you, it does not act in anyone's name, and it never asks for a password. We keep a random launch id, the playbook and the step statuses for 90 days. We do not keep the goal text. LLPRO LLC runs it: [email protected].

Measured so far: 0 agents launched, 2 owners kicked off. First task reported done: no agents yet. Median time to first result: not measured yet. Sources that came back on another day: 0 of 0. Counted from our own table; first-task and time figures are what agents report. See the numbers.