A: Great questions — I’ll answer each one clearly and without marketing fluff.

1) Does Dialora provide an MCP server?

No. Dialora does not expose an MCP (Model Context Protocol) server today.

Dialora is not a model-hosting or agent-runtime platform in the sense MCP is designed for. We do not expose low-level model orchestration, token streaming, or context injection endpoints the way MCP servers do.

Instead, Dialora operates as an application-layer AI voice system:
• We manage the LLM orchestration
• We manage speech → reasoning → action → speech
• We manage telephony, latency, turn-taking, and reliability

So today:
• ❌ No MCP server
• ✅ High-level AI agent orchestration handled internally

2) Does Dialora have an OpenAPI? Is it on the roadmap?

Current state

Yes — Dialora already has API access, but it is:
• Event + action–based
• Focused on practical integrations, not raw model control

Current API capabilities include:
• Triggering calls
• Receiving call events (completed, transferred, failed, etc.)
• Fetching call metadata, transcripts, summaries
• Creating leads / pushing data to CRMs
• Webhooks for real-time automation

This is exposed via:
• Native API
• Webhooks
• Make / Zapier / Pabbly / n8n-style connectors (Tier 3+)

Roadmap

A public, fully documented OpenAPI spec is:
• ✅ Planned
• 🎯 Focused on agent control, analytics, and orchestration
• ⏳ Rolling out progressively (not a single “big bang” release)

We’re prioritizing:
• Agent lifecycle control
• Call execution APIs
• Usage + billing visibility
• Enterprise automation use cases

Exact date depends on stability milestones, but this is actively being built, not a “someday maybe”.

3) Can Dialora AI voice agents work with other AI agents?

Yes — at the workflow and event level (not at the raw model level).

Dialora agents can:
• Trigger other AI agents
• Be triggered by other AI agents
• Exchange structured data with them

How this works in practice

Dialora emits events, such as:
• Call started
• User intent detected
• Qualification complete
• Appointment booked
• Escalation required

You can use these events to:
• Call another AI agent (text, workflow, RAG, decision agent)
• Send data to an AI system that decides next steps
• Return instructions back to Dialora via webhook/API

Example architecture
• Dialora = voice interface + real-time interaction
• External AI agent = reasoning, planning, decisioning
• CRM / backend = state + persistence

This is typically implemented via:
• Webhooks
• API calls
• Automation platforms (Make, Zapier, n8n, Pabbly)
• Custom middleware

What Dialora does not do (by design)
• We do not let external agents take over real-time turn-taking mid-call
• We do not expose token-level streaming control

That’s intentional — it keeps calls stable, low-latency, and production-safe.

Summary (plain English)
• ❌ MCP server: No
• ✅ APIs & webhooks: Yes
• ✅ OpenAPI-style public spec: On the roadmap
• ✅ Can work with other AI agents: Yes, via events, APIs, and workflows
• 🎯 Dialora’s role: AI voice execution + orchestration, not raw model hosting

If you’re trying to build agent swarms, decision trees, or AI-driven operations, Dialora fits cleanly as the voice interface layer in that stack.

If you want, tell me what you’re trying to build (architecture-level), and I can tell you whether Dialora fits — and how to wire it correctly.