Flexible pricing for your needs
See every conversation your business has, get a straight read on what to fix, and let agents act only inside the rules you approve. You stay in control.
No card required
Free forever
Smart AI
$144 billed yearly · save $36
Deepest reasoning
$372 billed yearly · save $96
Compare every plan
Free Free forever | Starter $15 per month | Pro $39 per month | |
|---|---|---|---|
| Conversations & AI | |||
| Manta Chat | 10 / month | Unlimited | Unlimited |
| Premium model (Opus tier) | — | — | |
| Connected tools | 1 tool | Unlimited | Unlimited |
| Live call monitoring | Daily recap | ||
| 1000+ tasks with sources | Limited | ||
| Transcripts + recordings | — | ||
| Actions & automation | |||
| Book, follow up, and collect | — | ||
| Test calls + sandbox numbers | — | ||
| Approval rules + hand-off to a person | — | ||
| Multi-role routing | — | ||
| Account & support | |||
| Multi-location / team | — | — | |
| Priority support | Community | Priority | |
Build on the Manta brain
Pull the same agentic voice and call-handling engine into your own app, agent, or platform. Usage-based, with $25 in trial credits.
Calculate your costs based on your usage needs
Production
Recommended for your usage
- 10k agent actions / month
- Cancel anytime
- Yearly billing, save 20%
API questions, answered.
Building something specific? Talk to our engineering team
How do trial credits work?
A new API key ships with $25 in credits. That covers roughly 150 agent actions or several thousand individual tool calls, depending on how you use the surface. No card required to start, and no auto-conversion to paid.
When do I move from Trial to Production?
When you're ready to ship to real traffic. Production is capacity-aware: we look at your actual call shape and volume and quote accordingly. Volume discounts kick in as you scale; we don't play the per-call gotcha game.
Can I bring my own LLM?
Yes. Register Manta as an MCP server and pull the tool surface into any Claude, GPT, or open-model agent. You're paying for the tools and the orchestration, not the model.
Can the API place real calls?
Sandbox numbers and transcript analytics ship in the public surface today. Placing real calls is opt-in and gated per number. When you're ready we wire it up for your specific phone system, with sandbox routing during pilot.
What's in the Enterprise tier?
Dedicated rate allocation, a named on-call engineer, white-label SDK and docs, compliance and audit hooks, and an SLA. Procurement-friendly contract, MSA, and security review.
How do I get support?
Trial: community + docs. Production: email and a shared Slack channel. Enterprise: a named engineer who knows your integration and joins your incidents.