Skip to content

Hyponema Docs

Hyponema is a platform for AI agents that keep context across conversations.

Use these docs to create workspaces, configure personas, connect speech and model providers, launch web or telephony sessions, attach tools, and inspect what happened after each turn.

Hyponema is currently in private beta. Some features require an enabled workspace, provider credentials, or a Hyponema-issued API key.

  • Follow the quickstart to create your first agent and session.
  • Read the concepts if you are mapping Hyponema to your product model.
  • Use the guides when wiring the widget, text chat, voice WebSocket, tools, telephony, or observability.
  • Use the API reference for supported public integration flows.
  • Persistent memory for repeat conversations.
  • Versioned personas, agents, and runtime guardrails.
  • Vendor-neutral STT, LLM, and TTS configuration with cascading fallbacks.
  • Web widget, text chat (SSE), voice WebSocket, and telephony entry points.
  • Workspace tools (HTTP and MCP), knowledge base, dynamic variables, and listening profiles.
  • Outbound calling, scheduled callbacks, and webhook delivery.
  • Trace-level observability, evaluations, and post-session extraction runners.
GoalStart with
Try the smallest integrationQuickstart
Understand what an agent isAgents
Model users and memoryUsers and Memory
Add Hyponema to a websiteWeb widget
Stream text responsesText chat SSE
Build a voice clientVoice WebSocket
Let the agent act on your productTools
Give the agent reference materialKnowledge base
Place outbound callsOutbound and scheduling
Review production behaviorObservability