A clear contract. A more dependable conversation.
Your agent’s behavior deserves more than a long prompt. Bring goals, tools, knowledge, and safeguards together in a structured Agent Spec.
Describe & review
Make the instructions your own
Test a conversation
Hear it before you launch
Publish a version
Improve from real outcomes
Business brief
Your goals and caller context.
Agent Spec
Structured, validated configuration.
Browser test
Hear and inspect the experience.
Published version
A traceable starting point for operation.
Start with the work to be done
Describe the business, the calls you want to handle, and the moments that need a human. This brief becomes the starting point for a structured agent configuration.
Inspect the actual contract
Agent Spec JSON makes the agent’s configuration reviewable. Inspect goals, tool definitions, conversation rules, and compliance settings before you publish.
- Runtime validation through shared schemas
- Provider-neutral agent configuration
- Versioned changes instead of an untracked prompt
Test the paths that matter
Use browser calls to test normal conversations, missing information, interruptions, and handoffs. Review transcripts and available events, then refine the configuration.
Publish something you can trace
Deploy a known version and review the conversations it produces. Keep critical changes in the audit trail and use observed outcomes to guide the next revision.
Hear the difference.
Then build your own.
Start with a free workspace. Review your agent. Test before you publish.