Voice agent · Custom HTTP guide · Conversation-initiation webhook + post-call webhook

Northsun + ElevenLabs

ElevenLabs' documented conversation-initiation webhook (its docs also call this "Twilio personalization" or the "pre-call webhook") and post-call webhook cover the inbound-phone lifecycle for an ElevenLabs Agent. This guide maps those events to Mnemix through your server; it is a custom HTTP guide, not a native or certified adapter.

Request Northsun accessAPI base URL ↗Read the full guide
What
Mnemix is the memory and enrichment layer for AI agents. Governed context intelligence is the differentiator.
Who
For teams building voice, chat, coding, and workflow agents that need persistent, inspectable context.
Price
Hobby is $0. Paid access is Contact sales.
How
Use the frozen caller REST contract for recall, durable call write-back, and read-only caller lookup. Use authenticated MCP only for the broader beta context tools enabled and proven for your tenant.
For agents
On an inbound call your voice agent receives the caller profile, this tenant's prior-call memory, and available Trestle and Twilio Lookup enrichment as contextual evidence before its first sentence; after hangup the call is submitted and committed durably before the response returns, visible on the very next read.

What Northsun adds to ElevenLabs

How to connect ElevenLabs to Northsun

  1. 1

    Configure the conversation-initiation webhook

    Follow ElevenLabs' current dashboard configuration and enable the system-prompt and first-message overrides your response depends on.

  2. 2

    Return the client_data envelope

    Call Mnemix, then return conversation_initiation_client_data with only the dynamic_variables your agent's prompt uses.

  3. 3

    Configure the post-call webhook

    Verify the elevenlabs-signature HMAC and map the current post_call_transcription payload into Mnemix's calls/end schema.

  4. 4

    Test both call directions

    Inbound and outbound/Register Call setups are different paths. Record a controlled proof for each direction you support.

Custom HTTP pattern

Adapt this boundary to the current vendor payload and verify it with controlled events before production.

async function callMnemix(path: string, body: unknown) {
  const response = await fetch(`https://mcp.mnemix.ai${path}`, {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.MNEMIX_API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify(body),
  });

  if (!response.ok) {
    throw new Error(`Mnemix request failed: ${response.status}`);
  }

  return response.status === 204 ? undefined : response.json();
}

export async function POST(request: Request) {
  // Conversation-initiation webhook: verify your configured shared-secret header first.
  const payload = await request.json();
  const packet = await callMnemix("/v1/recall_and_enrich", {
    phone_number: payload.caller_id,
    trigger: "answered",
    session_id: payload.call_sid,
  });

  return Response.json({
    type: "conversation_initiation_client_data",
    dynamic_variables: {
      caller_name: String(packet.caller?.name ?? "there"),
      caller_summary: String(packet.memory?.summary ?? ""),
    },
  });
}

// In your verified post-call webhook handler (elevenlabs-signature HMAC):
await callMnemix("/v1/calls/end", normalizedCallEnd(postCallPayload));

FAQ

Is this a native ElevenLabs integration?
No. It is a custom HTTP mapping maintained in your own application against ElevenLabs' documented webhook contract.
Does this cover the web widget?
No. The web widget has no phone number, and Mnemix recall is phone-keyed, so it is out of scope for this guide.
Does one successful webhook prove write-back?
Yes for the interaction itself: a Mnemix 204 confirms it is already committed durably, and the next recall for that caller sees it immediately. It does not prove your ElevenLabs mapping is correct — verify with a later recall in your own integration.

Last updated: .