Sandbox key · synthesis + TTS

Broadcast-grade audio APIsShip in days, not quarters

Synthesis, TTS, provenance, translate, and briefings on a single /v1/ surface — bearer keys, per-tenant metering, and predictable latency targets.

Start your project Dashboard Documentation

1k free calls/mo · sandbox & live keys · no card

POST/v1/synthesiscURL

curl -X POST https://api.narvo.news/v1/synthesis \
  -H "Authorization: Bearer nrv_sk_live_a3f8c2..." \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/story",
    "language": "en",
    "include_audio": true
  }'

200 OK2.28s · typicalJSON

{
  "story_id": "a3f8c2e1-7b9d-4e2a-8f1c-6d5a2b1e9c8f",
  "narrative": "Markets opened higher after guidance beat consensus...",
  "key_takeaways": [
    "Guidance raised for Q3",
    "Margins stable YoY"
  ],
  "source_attribution": [
    "Reuters",
    "Bloomberg"
  ],
  "language": "en",
  "audio_url": "https://cdn.narvo.news/audio/a3f8c2/en/emma.mp3",
  "provenance": {
    "source_count": 3,
    "corroboration_level": "strong_consensus",
    "consensus_score": 0.89
  },
  "latency_ms": 2280,
  "request_id": "req_01HN8VXQ4K2B7M9P3R5T6W8Y0Z"
}

Six engines, one versioned REST contract

Bearer keys, request_id on responses, optional provenance — built for broadcast-scale workloads.

POST/v1/synthesis

Synthesis

URL or text → narrative, takeaways, sources, optional audio.

API reference

POST/v1/tts/generate

Broadcast TTS

African voices (YarnGPT + fallback). MP3 or streaming Opus.

emmaiderazainabosagiewura

Voices / formats

POST/v1/provenance/analyze

Provenance Tag

Consensus scores from fact-checkers and Tier-1 Nigerian sources (RAG).

Africa Check, Dubawa, PolitiFact
Tier-1 NG outlets
Consensus inline

How scoring works

POST/v1/translate/narrate

Translate & Narrate

Journalistic register for African languages. Text or text + audio.

Supported pairs

GET/v1/briefing/latest

Morning Briefing

Daily bundle: audio, segments, provenance — ready to embed.

Briefing payload

WSwss://ws.narvo.news/v1/stream

Realtime / WebSocket

Breaking fan-out for widgets, dashboards, telco alerts (roadmap).

Roadmap

One engine or all six.

Broadcast APIs, one metered platform.

Live pipeline

0+: Curated feeds
0: Languages (core tier)
0s: P95 synthesis

Teams shipping audio news

From publishers to telcos — same versioned contract, metering, and audit trail.

Built for

Publishers
·
Newsrooms / radio
·
Telcos
·
PR / Gov
·
NGOs / research
·
Podcasts

Built for production

What operators ship first when they wire Narvo in.

Global publisher

Cut audio packaging time from days to hours: one contract, one meter, and request_id on every response for downstream QA.

4×faster to air

Broadcast SaaS

We white-label synthesis + provenance for partner stations. Tenant-scoped keys and usage exports made finance reviews painless.

99.6%uptime (90d)

Telco media unit

Cached briefings for IVR and USSD menus; P95 synthesis stayed inside our SLA window under peak traffic.

P95<3s synthesis

First call in minutes

Create a sandbox key in the dashboard, send a bearer header, then call /v1/synthesis. The playground mirrors production fixtures.

Playground Dashboard More examples

terminal

curl -sS -X POST https://api.narvo.news/v1/synthesis \
  -H "Authorization: Bearer $NARVO_SK" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com/story","language":"en"}'

Guarantees

What paid newsrooms expect on day one.

Multi-tenancy

Tenant scope comes from the API key — not client-supplied IDs alone.

Per-tenant keys
Quota isolation
Audit logs

Usage metering

Each 200 OK logs tenant, endpoint, and tokens atomically.

Atomic writes
Per-language splits
Exports

Rate limits

Sliding windows per tier; 429s include retry guidance.

Synth 10/min
TTS 20/min
Provenance 60/min

SLA enforcement

Tier uptime targets; credits if a window misses.

Publisher 99.5%
Newsroom 99.7%
Enterprise 99.9%

Ship audio news
at platform scale

1k free calls/mo · sandbox & live keys · no card. Developer tier stays free while you ship.