Prepare your agent to work with (and against) other agents.
Put your agent in simulations with other agents that pursue their own goals and react to its decisions. Test how it handles cooperation, conflict, and unexpected responses.
Particles form the Simulary logo, then drift apart and settle back. Nearby points respond to the pointer. This is decorative animation.
Get an email
when we launch.
Agents will do business
with other agents.
As people delegate more work to AI, their agents will need to coordinate across teams and businesses. A buyer’s agent might negotiate a price with a supplier’s agent. A scheduling agent might arrange a meeting with someone else’s assistant.
You won’t control every agent yours encounters. We’re building Simulary to help you test those interactions before you depend on them.
Test agents across different kinds of work.
Purchasing agents
See how your agent negotiates price and delivery with supplier agents. Vary their stock and priorities, and check whether your agent stays within budget.
Scheduling agents
Test how your agent finds a meeting time with other assistants. Change their availability, add a scheduling conflict, or move the meeting at the last minute.
Customer support agents
Test how your agent resolves a request with another agent’s help. Try missing information, failed handoffs, or disagreements over who should act.
Work coordination agents
Test how your agent assigns work and checks the result. See what it does when another agent misses a deadline or returns incomplete work.
Test how your agent
interacts with others.
Run your agent in a scenario with other agents. Give each one a role and a goal, and set the rules they follow. See how your agent responds to their decisions.
Define the scenario.
Choose the task, the other agents, and what each agent can see and do. Test cooperation, competition, or a mix of both.
Inspect the decisions.
Read each message and action in order. See what information your agent had, what it chose to do, and how the others responded.
Compare agent versions.
Change your agent’s instructions or code, then rerun the same scenario. Compare the results from the same starting conditions.
Follow each
agent’s decisions.
Our approach
We record what each agent could see, what it did, and how the others responded. Compare agent versions in the same scenario. Use those results to choose what to test next.
Use a browser or
your coding assistant.
We’re building both paths so you can set up simulations and review results, whether or not you write code.
In your browser
Describe the task and connect your agent. We’ll walk you through setup without asking you to write code.
With your coding assistant
Ask your coding assistant to set up tests, run simulations, and review results alongside your agent’s code.
Simulary is coming soon.
We’re preparing the first release.