Building cosmic-grade ventures that executives can trust without a committee.

AngelList · Fundraising

SmartQueue

A government-first intake and service-queue platform built for procurement teams, compliance requirements, and transparent public reporting. It standardizes service requests, assigns ownership, and keeps every step auditable for FOI, SLA, and vendor-review workflows.

Designed for city and agency contracts, SmartQueue emphasizes clear escalation paths, documented service levels, and analytics that translate into budget justification. It is structured to support stakeholder visibility without exposing sensitive citizen or vendor data.

AngelList · Fundraising

AI Concierge

Support intelligence, shipped as an API.

Grounded answers from your docs. Clear escalation when unsure. Built for teams that want control.

Powerful capabilities, simple API

Everything you need to build intelligent support, with the control and transparency developers expect. One endpoint, endless possibilities.

Terminal

curl https://concierge-edge-proxy-1024642568824.us-central1.run.app/v1/answer \
  -H "Authorization: Bearer $ZARD_API_KEY" \
  -d '{"query": "What is your refund policy?"}'

{
  "answer": "Our refund policy allows returns within 30 days...",
  "confidence": 0.94,
  "citations": [{"snippet_id": "snp_x7f2..."}],
  "action": "answer",
  "request_id": "req_9k3m..."
}

Core capabilities

Every feature designed for production reliability.

• Knowledge grounded answers All responses are generated strictly from your knowledge base. No hallucinations, no guessing, just accurate answers with source citations.
• Confidence + citations Every answer includes a confidence score (0-1) and references to the exact snippets used. Debug and audit with full transparency.
• Escalation decision When confidence drops below your threshold, returns action=escalate instead of guessing. Route uncertain queries to human support automatically.
• Webhooks for handoff Configure webhooks to get instant notifications when queries need escalation. Integrate with Slack, your ticketing system, or custom workflows.
• Usage metering Track every API call with detailed metrics. Monitor usage, set rate limits, and get cost breakdowns by workspace or API key.
• Org and key isolation Complete workspace isolation ensures data never crosses boundaries. Manage multiple API keys with different permissions per environment.
• Drop-in chatbot widget Copy-paste a React chatbot component into any app. Zero configuration, just set your API key and you are live in minutes.

Developer payloads

{
  "answer": "...",
  "citations": [
    {"snippet_id": "snp_x7f2"}
  ]
}

{
  "confidence": 0.94,
  "threshold": 0.70,
  "action": "answer"
}

{
  "action": "escalate",
  "confidence": 0.32,
  "reason": "low_confidence"
}

POST /your-webhook
{
  "event": "escalation",
  "query": "...",
  "request_id": "req_..."
}

zard_sk_prod_... -> prod workspace
zard_sk_dev_...  -> dev workspace
zard_sk_test_... -> test workspace

import AgenticChatbot from "./AgenticChatbot";

<AgenticChatbot />

Simple integration

How it works. Three steps to intelligent support.

• 01 Upload snippets Policies, FAQs, documentation.
• 02 Call /v1/answer One simple API endpoint.
• 03 Get answers or escalate Route intelligently to your stack.

Enterprise ready

Developer trust, built-in. Every response includes traceability. Every request is logged. Everything you need to debug, audit, and scale with confidence.

• Request IDs on every responseEnd-to-end traceability.
• Strict JSON outputStable schemas for your parsers.
• Rate limits + usage countersVisible controls per workspace.
• Keys never stored in plain textSafer operational hygiene.

Response headers

x-request-id: req_9k3m4f7x2n
x-confidence: 0.94
x-latency-ms: 127
x-ratelimit-remaining: 9847
x-ratelimit-reset: 1703980800
content-type: application/json

Use cases

What you build with it. From customer support to internal tools.

• In-app support widget Embed intelligent support directly in your product without breaking flow.
• Documentation assistant Help users navigate complex docs, API references, and guides.
• Internal support bot Answer employee questions about policies, procedures, and internal tools.
• Helpdesk deflection layer Reduce support ticket volume by 40-70% before they reach your team.