AI voice kit for Kenya.
Get a Kenyan number for KES 100/month, run managed AI agents and voice broadcasts, and control every call from one API — billed per second in KES. More African markets are on the way.
# Deploy an AI agent on your number
curl -X PUT https://api.sautikit.com/v1/numbers/$NUMBER_ID/routing \
-H "Authorization: Bearer $SAUTIKIT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"inbound_agent": {
"enabled": true,
"agent_id": "9d2b1f53-4c1e-4f7a-9b0e-2f8c1d6a7e21",
"mode": "always",
"handover_callback_url": "https://your-app.example.com/handover"
}
}'Managed AI calls
AI agents for inbound. Broadcasts for outbound.
Choose a managed agent for calls to your number or a broadcast for calls to your list. Both run from the same prepaid wallet.
AI usage adds KES 4.00/min to standard call rates, metered per second. The free tier supports one concurrent AI call.
Managed inbound AI agents
Publish an agent on your SautiKit number to answer around the clock or only after hours. It holds a natural conversation, uses your systems, and hands calls to your team with context.
- Answers 24/7, or only outside your business hours
- Calls your systems mid-call with signed HTTP tools
- Briefs your team on handover — reason and summary included
Outbound AI broadcasts
Schedule an agent to call a contact list. Each person gets a personalized, interruptible conversation within the calling windows you set, followed by a report for every call.
- Personalized per contact with template variables
- Calling windows, backoff, and retry policies
- Detects voicemail and retries later
Stream live call audio to your AI.
The <Stream> verb forks live call audio to your WebSocket and returns audio to the caller. Connect your own model when you need direct control of the voice experience.
- Full-duplex WebSocket audioFork live PCM frames to your socket and stream audio straight back into the call.
- Use the model you choosePoint call audio at Gemini Live, OpenAI, or a self-hosted model.
- Analyze calls while they happenTap either leg for live transcription, sentiment, or call scoring.
Available today via XML from your voice webhook. Native JSON stream action is on the roadmap.
<Response>
<Stream
name="ai-agent"
url="wss://your-app.example.com/audio"
track="both_tracks"
connect="true"
outputSamplingRate="16000"
bidirectionalSamplingRate="16000"
statusCallback="https://your-app.example.com/stream-status" />
</Response>What you get
The building blocks for AI voice in Kenya.
Provision, route, automate, record, stream, and measure every call from one API.
Managed AI call agents
Publish an agent on your number and hand calls to your team with a summary.
Voice broadcast campaigns
Personalized AI calls at scale with calling windows, retries, and a CSV report.
Kenyan phone numbers by API
Provision local numbers in seconds and assign them to a call flow.
Call routing and control
Place, receive, forward, and branch calls without middleware.
Per-second KES billing
Top up a prepaid wallet and pay only for the seconds you use.
Webhooks with automatic retries
Real-time callbacks for every call event, retried with backoff.
Call recording and storage
Record any leg with one flag; the first 1 GB of storage is free.
Call logs, costs, and events
Query duration, status, cost, and DTMF events by API or console.
Build with AI
Give your coding agent control of your phone system.
Connect Claude Code, Cursor, or VS Code to the hosted SautiKit MCP server — 45 workspace tools for numbers, calls, agents, and broadcasts.
Pricing
Per-second pricing in KES.
Full pricing →Make your first API-controlled call today.
Create an account, top up your wallet, and make your first call in minutes. No contracts or credit checks.