Skip to main content
Roadmap

What ships when.

Frontbell is live in production, onboarding trade businesses in a controlled beta, with open availability rolling out through Fall 2026. Everything after that is contingent on the launch product working — we'll ship the next thing when the previous thing is real.

Now

2026 — launch year

Q1 2026 · Done

Voice pipeline + CE-OS engine

Custom WebRTC transport, full-duplex orchestrator, streaming ASR/LLM/TTS, state-machine workflow engine. CE-OS patent filing submitted (USPTO).

Q1–Q2 2026 · Done

Per-tenant telephony & SMS compliance

Twilio Subaccount-isolated tenants, 10DLC brand registration flow, A2P consent capture, recording disclosure.

Jul 2026 · Shipped

Private beta live — painting, cleaning, carpentry/small remodels

Frontbell is in production: web + mobile app with leads pipeline, line-item estimating with production rates, scheduling and dispatch, invoicing, photos, and the Brion assistant. Voice front desk rolls out per tenant as business lines are provisioned.

Fall 2026 · Open availability

Open availability — all three trades

General availability once the beta cohort has exercised the system; phone and SMS lines provisioned per tenant as carrier registration clears.

Late 2026

Stripe-powered online estimate acceptance & payments

Customer signs the estimate, pays the deposit, and the work is on the calendar — without the owner picking up the phone again.

Q4 2026 – Q1 2027

Per-tenant analytics & lead-temperature scoring

Pattern engine post-launch: per-tenant lead-quality scoring trained on call outcomes. Customer health metrics for the owner dashboard.

Next

2027 — depth and economics

2027 H1

Voice cloning (in-app guided owner capture)

Mobile-guided high-quality voice capture, per-tenant cloned outbound voice with consent disclosure. Owner-only outbound calls at first.

2027 H1

Self-hosted ASR + intent inference (NVIDIA evaluation)

Benchmark and ship self-hosted ASR (NVIDIA Riva Parakeet candidate) and small fine-tuned intent/routing models on TensorRT-LLM. Goal: 30–40% per-call cost reduction and tighter latency budget.

2027 H2

Additional trade verticals

Expand to additional trades once the first three are stable and unit economics are proven. Each new trade is a configuration, not a separate codebase.

2027 H2

Outbound voice (consented)

Outbound follow-ups, appointment confirmations, and review requests — strictly consent-gated, owner-attributed. Carrier-side compliance is the gating constraint, not the engineering.

Later

2028+ — what we're not promising yet

Anything past 2027 depends on what works. Some directions we're tracking but not committing to:

We'd rather under-promise the long roadmap and over-deliver the short one. The beta is live today; open availability through Fall 2026 is the only forward date that's load-bearing.

Want updates?

Email us — we'll send a short note when each milestone ships.

hello@vaxio.ai