On August 21, 2025, Ibadan-based software developer Dára Sobaloju published a post on X that was meant as a simple product reflection. He was not announcing a launched product, nor did he have a venture team behind him; he wanted to gauge whether a long-standing frustration in physical worship spaces was worth solving.
His proposition was simple: an artificial intelligence agent that listens to live sermons and projects matching Bible verses onto church screens in real time as the pastor speaks, paraphrases, or spontaneously quotes scripture.
The response from pastors, media teams, and churchgoers across the globe was immediate. Live presentation teams in large congregations routinely struggle to keep pace with spontaneous preaching, causing delays that disorient speakers and break the congregation’s focus. Recognizing the immediate demand, Sobaloju began building. Six months later, PewBeam AI was live as a full-featured desktop application across Windows and macOS.
Engineering for the Edge: Speed, Context, and Offline Independence
Building a real-time speech processing tool for live environments presents complex technical constraints. Church auditoriums introduce acoustic resonance, background music, and variable microphone quality. Furthermore, relying on active internet connections during large gatherings is notoriously unreliable.
To solve these challenges, PewBeam was architected around three foundational principles:
-
Sub-100ms Latency: While early iterations processed audio with a 500-millisecond delay, the current architecture brings processing speed down to between 2 and 100 milliseconds (averaging under 80 milliseconds). The verse appears on screen virtually as the preacher speaks it.
-
Contextual Intent Matching: Rather than relying solely on rigid keyword searches, the underlying NLP models recognize paraphrases, partial quotes, and thematic intent to surface the precise passage.
-
Offline-First Resilience: The application operates entirely without an active internet connection, utilizing local speech recognition and built-in noise cancellation.
“Anything that can work in Nigeria can work outside Nigeria. It’s a harsh environment.”
— Dára Sobaloju, Founder of PewBeam AI
Product Performance and Operating Metrics
PewBeam combines real-time speech-to-text with flexible hardware integration, enabling media teams to output displays via HDMI, in-app previews, or NDI network streaming directly into broadcast setups.
┌───────────────────────────────────────────────────────────────────────────┐
│ PEWBEAM AI METRICS & CAPABILITIES │
├───────────────────────────────┬───────────────────────────────────────────┤
│ Processing Latency │ 2ms – 100ms (Sub-80ms Average) │
│ Global Deployment Footprint │ 7,100+ Churches across 20+ Countries │
│ Domestic Penetration │ ~70% of Nigeria's Largest Congregations │
│ Customer Retention Rate │ ~80% Active Subscription Retention │
│ Revenue Distribution │ 60% Diaspora / 40% Domestic Paid Ratio │
│ Supported Display Protocols │ HDMI, In-App Preview, NDI Streaming │
└───────────────────────────────┴───────────────────────────────────────────┘
Challenging Monetization Assumptions in African SaaS
A common assumption among venture capital investors is that local institutions in emerging markets, particularly religious organizations, resist recurring software subscriptions. PewBeam’s traction directly challenges that narrative.
The platform operates a freemium model featuring a starter tier with up to 40 minutes of weekly transcription, alongside paid subscription tiers ranging from $14 to $30 per month for expanded usage, multi-device sync, and advanced streaming output.
Currently, the revenue split between diaspora congregations and domestic Nigerian churches sits at 60/40. The fact that 40% of paying subscribers are local Nigerian institutions highlights a clear market reality: emerging market customers will pay for software when it directly solves an acute operational bottleneck.
Supporting this paid adoption is an agile development cadence. In one instance, when a customer in the United Kingdom requested an unreleased feature during a live service, the PewBeam team built, tested, and shipped the update in under 30 minutes for immediate use.
The Broader Vision: Building Church-First Infrastructure
Historically, religious institutions have adapted secular tools—such as office presentation software and video conferencing suites—for liturgical use. PewBeam is attempting to invert this relationship by engineering high-performance communication software specifically for worship environments that can later scale into secular enterprise applications.
The team has already begun expanding its product suite:
-
PewNote Integration: A newly introduced interactive sermon layer that generates searchable, interactive transcripts allowing congregants to study and query sermon notes between weekly services.
-
Domain-Specific AI Models: Long-term plans include training proprietary AI models on scriptural context and theological nuances, creating an API foundation that third-party faith-based developers can build upon.
-
Global Footprint: Targeting expansion into at least 120 countries over the next five years.
The African FaithTech Movement
PewBeam represents part of a broader wave of African developers building localized, high-utility software for some of the world’s largest physical gatherings. Similar innovations—such as Tolulope Adeniyi’s Spetra verse detection engine, Olanrewaju Taiwo’s Meno context-aware Bible platform, and institutional deployments like Deeper Life’s Ask Kumuyi AI assistant—signal that African engineers are leading the digital transformation of religious infrastructure.
As religious institutions globally transition deeper into hybrid and tech-enabled services, tools built under the network and environmental constraints of markets like Ibadan are proving to be among the most resilient software architectures in the world.
