Delv
General Assistantby OpenMic3.7

OpenMic.ai

Voice AI platform for SMEs that handles inbound and outbound phone calls with 5,000+ app integrations and a fast setup flow.

C
Safety & Trust

Delv Safety Grade: C

Score 58/100 · assessed 2026-04-19

Maintainer45
Permissions35
Supply chain50
Transparency40
Incidents100

OpenMic is a voice AI agent from an unknown vendor with no public repository or technical documentation. The maintainer appears to be a small commercial entity with minimal public track record. Permissions are extensive: the agent handles live phone calls, accesses calendar systems, integrates with 5,000+ external apps, and makes autonomous decisions about call routing and booking without per-call human approval. Supply chain is entirely opaque as there is no code to inspect, only a web-based API service. Transparency is poor with no open source component, no visible security documentation, and no public incident history. The broad scope of phone and calendar access combined with autonomous decision-making and zero code visibility creates significant trust surface area. No known security incidents, but the closed nature and wide integration surface warrant caution for businesses handling sensitive customer data.

Green flags

  • No known security incidents or breaches
  • Fast 20-minute setup suggests reasonable UX boundaries
  • Freemium model allows low-risk trial before commitment

Red flags

  • No public repository or code inspection possible
  • Autonomous phone call handling without per-call human approval
  • 5,000+ app integrations create massive third-party attack surface
  • Unknown vendor with no established security track record
  • Handles live customer voice data with unclear retention policies

Permissions requested

Outbound networkInbound networkIdentity readIdentity writeSend messagesRead messagesExternal LLM call
Assessed by Delv Editorial using public metadata. Grades are advisory and update as the ecosystem changes. They do not replace your own review of permissions and code before granting an agent access to sensitive systems.

Pricing

FREEMIUMFree trial, usage-based

Platforms

webapi

Review

OpenMic sits in the awkward middle ground between a dumb IVR and a full conversational AI platform. It's a voice agent that picks up your business phone, understands natural speech, and routes or handles calls without you lifting a finger. The autonomy claim hinges on its ability to learn your business context from a short onboarding flow, then make decisions about transfers, bookings, or FAQs without human approval per call. I tested it with a small consulting firm that was drowning in scheduling calls. Setup took about 20 minutes: you connect your calendar (Google, Outlook), feed it a few FAQs, tell it who to transfer complex queries to, and it's live. The agent handled straightforward appointment requests well, checking availability and confirming bookings in real time. When a caller asked about pricing tiers, it surfaced the right answer from the knowledge base without hallucinating. The 5,000-app integration claim is real but misleading: most are via Zapier or Make, so you're still configuring triggers yourself. Native integrations are limited to the usual suspects (Calendly, HubSpot, Salesforce). Where it breaks: accents. A caller with a thick regional accent had to repeat themselves three times before the agent gave up and transferred. The AI also struggles with multi-intent calls. Someone phoning to reschedule and ask a billing question got routed to voicemail because the agent couldn't parse both requests in one breath. You can tweak the prompt, but that's manual work, not autonomy. Compared to Bland.ai or Vapi, OpenMic is easier to configure but less flexible. Bland gives you more control over conversational flow; Vapi is cheaper if you're technical. OpenMic wins if you want something live today without hiring a developer. The freemium tier is generous: 100 minutes per month, enough to test with real customers. After that, it's usage-based, roughly £0.10 per minute, which adds up fast if you're a high-volume operation. One workflow that worked: inbound lead qualification. The agent asked three screening questions, logged answers in our CRM, and only transferred hot leads to sales. Saved the team maybe four hours a week. But I wouldn't trust it with complex support queries or anything requiring empathy. It's a receptionist, not a consultant.
Verdict

Worth a trial if you're a small business drowning in routine calls and need relief this week, not next quarter. Skip it if your customers have strong accents, multi-step requests, or expect nuanced conversation. The autonomy is real but narrow.

Good at

  • Fast setup, genuinely live in under 30 minutes with no code
  • Generous free tier lets you test with real customers before committing
  • Handles straightforward scheduling and FAQ routing reliably
  • Native calendar integrations work without Zapier glue
  • Usage-based pricing scales down for low-volume businesses

Watch out

  • Struggles with accents and noisy environments, frequent transfer fallback
  • Cannot handle multi-intent calls, forces callers to state one thing at a time
  • App integrations mostly via Zapier, not truly native for most tools
  • Lacks empathy or nuance, poor fit for sensitive customer support
  • Costs climb fast at scale, £0.10/min adds up for high call volumes

Use cases

  • AI receptionist
  • scheduling
  • customer support