What the Knowlarity Startup Program Gives You
Knowlarity offers up to $3,000 in credits toward its cloud telephony platform, which covers the meter that actually runs the bill: connected minutes on virtual numbers, IVR flows, call recording and outbound dialing across the public phone network.
AI Perks tracks it alongside $7.7M in credits across 194 companies.
The boundary is worth naming early. Knowlarity sells carrier-grade voice plumbing, so the credit absorbs telecom charges: per-minute call costs, monthly number rental, IVR time, dialer legs. It does not absorb the CRM you route calls into, the storage your recordings pile up in, or the speech and model bills behind any voice bot you put on top. The current amount and terms are listed on getaiperks.com.

What Knowlarity Is Actually For
Knowlarity is a cloud telephony and CPaaS platform built for India and neighbouring emerging markets, selling virtual numbers, IVR, call tracking, click-to-call, missed-call flows and hosted contact centre seats as an API and a dashboard instead of on-premise hardware. It has operated since 2009 and was acquired by Gupshup in a deal reported at $100M.
The relevant comparison is not Knowlarity against another vendor. It is whether your product needs the phone network at all, and if so, whether it needs it where Knowlarity is strongest.
Plenty of teams reach for telephony when in-app chat or push would do. The phone network earns its cost in three situations:
- Your customers are not in your app. Field agents, delivery partners, walk-in leads, first-time buyers responding to an offline ad. A number on a billboard converts people who will never install anything.
- You need reach independent of data connectivity. A voice call lands on a feature phone with no data plan. Where that is a meaningful share of demand, it is coverage, not nostalgia.
- Regulation or trust demands a real number. Masked numbers between two parties, verified callbacks, recorded consent.
Geography is the sharper filter. Cloud telephony is intensely local: numbering rules, carrier interconnects, consent registries and per-minute economics differ by country, and a vendor that is excellent in one region is mediocre in another. Knowlarity's centre of gravity is India and South and Southeast Asia, and AI Perks lists North America-first programs separately.
How Cloud Telephony Pricing Behaves at Scale
Cloud telephony meters call legs, not conversations, and rounds each leg up to a fixed pulse. Those two mechanics are where telephony invoices diverge from the model in your spreadsheet, usually by a factor of two or more.
A "leg" is one connection the platform establishes. Bridging two people means two legs, both billed, for the same conversation:
| Call flow | Legs metered | Minutes billed for a 3-minute conversation |
|---|---|---|
| Inbound call answered in-browser | 1 | 3 |
| Inbound call forwarded to an agent's mobile | 2 | 6 |
| Click-to-call triggered from your CRM | 2 | 6 |
| Outbound dialer bridged to an agent | 2 | 6 |
| Warm transfer to a second agent | 3 | 9 |
| 30 seconds of IVR before pickup | 1, extended | 3.5 |
The second mechanic is pulse rounding. Indian cloud telephony is commonly billed in 30 or 60 second pulses rounded up, so a 9-second call and a 29-second call cost the same. For a product built on very short calls, OTP delivery, missed-call triggers, delivery confirmations, your unit cost is set by pulse count, not by talk time. Confirm the pulse length in your own contract.
The meters that make up a real invoice:
| Meter | What drives it | What makes it spike |
|---|---|---|
| Connected minutes | Legs multiplied by pulses, inbound and outbound priced differently | Forwarding every call to a mobile |
| Number rental | Monthly fee per virtual number | A number per campaign, never retired |
| Toll-free | Higher rental, and you pay the caller's minutes | Buying 1800 before demand justifies it |
| Concurrency | Simultaneous channels, often a contracted ceiling | Campaign blasts and daily peak hours |
| Agent seats | Per-seat licence for contact centre features | Seats provisioned for churned staff |
| Recording and storage | Recording is a feature, retention is a bill | Indefinite retention, no lifecycle policy |
| SMS and WhatsApp | Separate per-message meters plus registration | Treating them as included in voice |
| Voice bots and transcription | Billed by speech and model vendors | Assuming telephony credits cover AI |
Publicly quoted Indian rates sit roughly between 0.30 and 1.50 rupees per minute depending on direction and call type, with local virtual numbers commonly a few hundred rupees a month and toll-free numbers several times that. Treat those as orientation and verify against your own quote. The credit amount is on AI Perks; the multipliers above are yours to control.

What Knowlarity Credits Stack With
Telephony credits stack cleanly because the invoices genuinely do not overlap, and the product most teams are building on telephony right now, a voice agent that answers the phone, is a four-vendor pipeline where telephony is the cheapest leg.
One turn of an automated phone agent runs: caller, carrier network, transport, speech-to-text, model, text-to-speech, back out through the carrier. Separate companies bill each hop:
- Telephony credits cover the carrier leg, the $3,000 on this page
- Speech credits cover transcription and synthesis, metered per second of audio
- Model credits cover the LLM, usually the largest line in the pipeline
- Cloud credits cover orchestration, recording storage and egress
- Messaging credits cover the SMS and WhatsApp follow-up that voice campaigns almost always trigger
A team holding three of those five has funded a phone-based product through its first year on grants alone. Which programs combine and which quietly exclude each other is why AI Perks is maintained as a live list.
What Founders Get Wrong About Cloud Telephony Credits
The most expensive mistake is budgeting conversations instead of legs, because a click-to-call product bills exactly twice what a naive forecast predicts, every single call, forever.
Five patterns, in rough order of what they cost:
Counting one leg. "2,000 calls a day, four minutes each" reads as 8,000 minutes. If every call is bridged, it is 16,000. Nothing about the product changed, only the unit.
Ignoring pulse rounding on short calls. A verification call that connects for eleven seconds bills a full pulse. At high volume, shaving seconds is worthless and reducing call count is everything.
Buying toll-free too early. Toll-free means you pay for the caller's minutes as well as the rental. Until inbound volume is proven, a local virtual number does the same job for less.
Planning minutes but not concurrency. Telephony fails on simultaneous channels, not on monthly totals. A campaign that sends 50,000 SMS at 10am creates a callback spike at 10:05 that exceeds your channel ceiling, and callers get a busy tone. Minutes are a budget problem, concurrency is an outage.
Assuming the credit covers the bot. In an AI phone agent, the carrier leg is typically the smallest of four costs. Funding it and not funding inference solves the cheap quarter of the problem.
One more, less about money: your published number ends up on packaging, ads and invoices, so portability terms matter earlier than they feel like they should. And decide at 70% of credit consumed what your unsubsidised bill looks like.

Where Knowlarity Fits in a Communications Stack
Telephony is the layer every phone product sits on, so the useful question is not whether the credit is large but whether the call shape it subsidises is the one your product generates.
Three things decide that, all knowable before a minute is billed:
Call shape. Inbound calls answered in-browser stretch the credit furthest. Outbound dialing bridged to human agents burns it at roughly double the rate, because every conversation is two metered legs.
Geography. Cloud telephony is national, not global, and Knowlarity's coverage runs deepest in India and neighbouring markets. The communications category on getaiperks.com lists the regional alternatives.
Adjacent meters. Speech, model and messaging bills sit on top of the carrier bill and are usually larger than it.
One thing costs nothing to plan and a great deal to retrofit: commercial calling and messaging in most markets requires consent and sender registration, and that registration, not the credit, usually decides when a phone product can start dialing.
Frequently Asked Questions
How much is the Knowlarity startup program worth?
Up to $3,000 in credits toward Knowlarity cloud telephony. How far that goes depends almost entirely on call shape: inbound calls answered in-browser stretch it a long way, while bridged outbound dialing at scale burns it quickly. The current amount and terms are tracked at getaiperks.com.
Why is my telephony bill double what I forecast?
Because most flows bill two legs for one conversation. Click-to-call, outbound dialing and forwarding to an agent's mobile all require the platform to establish two connections, and both are metered. Add pulse rounding, which charges a full unit for a partial one, and the gap between forecast and invoice widens further.
Does Knowlarity work outside India?
Knowlarity's strength is India and neighbouring South and Southeast Asian markets, where its numbering, routing and compliance coverage is deepest. Cloud telephony is intensely regional, so a US-heavy calling product is usually better served by a North America-first provider. Communications programs for other regions are listed at getaiperks.com.
Do telephony credits cover an AI voice agent?
No. Knowlarity bills for the carrier leg only. Transcription, model tokens and voice synthesis come from three separate vendors and together usually cost more than the call itself. Plan to hold speech and model grants alongside the telephony grant rather than instead of it.
What is the fastest way to cut a cloud telephony bill?
Three changes, in order of impact: answer calls in-browser instead of forwarding to mobiles to drop from two legs to one, retire unused virtual numbers and campaign numbers, and set a retention policy on call recordings. Each is configuration rather than architecture.
Can I combine Knowlarity credits with other startup credits?
Yes, and communications credits stack well because the invoices do not overlap. Cloud credits cover orchestration and recording storage, model credits cover the LLM, speech credits cover transcription and synthesis. AI Perks tracks $7.7M in credits across 194 companies, including which of them conflict.
Ship the phone number. Let someone else pay for the first hundred thousand minutes.