Product
Build and run AI voice agents.
Orbitali gives developers and agencies a real-time voice runtime with prompts, tools, knowledge, transcripts, observability, and usage control. You own the logic. We run the agent.
// Orbitali streams typed events to your Server URL
export async function POST(req: Request) {
if (!verifySignature(req)) return new Response("Unauthorized", { status: 401 })
const { message } = await req.json()
switch (message.type) {
case "agent:tool-call":
const result = await runTool(message.toolCall)
return Response.json({ result })
case "agent:assistant-request":
return Response.json({ instructions, greeting })
}
}Runtime capabilities
Agent configuration
Identity, instructions, prompts, and greetings. Version and tune behavior without touching your backend.
Tool calls
Typed webhook contracts to your Server URL to run real-world actions during the call.
Context request
The assistant-request flow fetches customer-specific context before the agent responds.
Knowledge base
Support for policies, FAQs, and product information the agent can consult.
Transcripts and logs
Capture and review transcripts, tool-invocation logging, and usage metrics.
Voice channels
Inbound routing through customer-owned Telnyx and Twilio numbers, plus controlled outbound calling from eligible Telnyx numbers. More providers can be added when customers need them.
The flow of a conversation
You own the logic. We own the agent.
Orbitali doesn't replace your backend. Customer-specific business logic lives in your own API via the Server URL.
Your backend
- Business logic
- Customer data
- Integrations and CRM
- Rules and decisions
Orbitali
- Managed STT, LLM, and TTS models
- Real-time voice runtime
- Orchestration and tools
- Knowledge and prompts
- Logs and observability