dTelecom STT

Real-time speech-to-text API for AI agents. Pay with USDC via x402. No API keys, no accounts.

Built on dTelecom DePIN — a decentralized real-time communication network.

x402 99+ languages $0.005/min MCP

Why This Service

Most STT APIs feed raw audio directly to a model. We don't. We are an audio intelligence pipeline — a complete system for turning noisy real-world audio into accurate text.

Features

• Full audio intelligence pipeline — real-time voice activity detection on dedicated hardware, neural noise reduction, speech validation, intelligent audio trimming, and hallucination filtering. Cleaner input = better transcription = no garbage output.
• Dual-engine architecture — Parakeet-TDT (3-4x faster, 25 European languages) + Whisper (99+ languages, contextual prompting). Smart routing directs requests to the best available engine with automatic failover. No single point of failure.
• 99+ languages — 25 European languages via Parakeet-TDT (3-5% WER on top languages), 99+ via Whisper. See docs for per-language WER benchmarks.
• Agent-native access — no API keys, no accounts, no signup. Any AI agent with a USDC wallet (EVM or Solana) can discover, pay, and use this service autonomously via x402.
• Fair usage-based billing — clock pauses on disconnect, only connected streaming time is billed. Reconnect without losing paid time. No other STT provider offers this.
• Privacy — dedicated hardware, local inference. Nothing stored, no audio logs, no training on customer data. Encrypted in transit (TLS/WSS).
• MCP server — use from Claude Code, Claude Desktop, Cursor, and other AI assistants out of the box

Competitive Pricing

Premium quality below major cloud providers:

Pricing

Quick Start

Python: pip install dtelecom-stt

TypeScript: npm install @dtelecom/stt

MCP (Claude/Cursor): npm install -g @dtelecom/stt-mcp