dTelecom

799 posts

dTelecom banner
dTelecom

dTelecom

@dtelecom

Communication layer for AI Agents & Humans. Real-time Voice, Video & Voice AI. Pay-per-use via x402 and MPP. Backed by @SolanaFndn, @Google, @ElevenLabs, @peaq

Beigetreten Mart 2022
27 Folgt119.4K Follower
Angehefteter Tweet
dTelecom
dTelecom@dtelecom·
The $3.5T telecom industry is being rebuilt on-chain ⛓️ @dTelecom is the default DePIN layer for real-time voice, video, and AI. Backed by: @SolanaFndn, @Google, @peaq Traction: 50K+ users 35M+ mins processed The $2.6M Airdrop is now LIVE Join: rewards.dtelecom.org
English
20K
21.3K
26.3K
441.9K
dTelecom
dTelecom@dtelecom·
@mpp is a machine payment protocol built for software (not humans). It lets agents handle micropayments, recurring flows, and usage-based service access programmatically. At @dTelecom, we use @mpp on @tempo so agents can buy credits and pay on demand for WebRTC, STT, and TTS through the same gateway they already use with x402 on Solana and Base. No API keys. No account setup. Just pay and use. Try it below ↓
English
7
11
73
2.2K
dTelecom
dTelecom@dtelecom·
The old software payment flow was built for humans: Signup ↓ Link a card ↓ Choose a plan ↓ Get the API key ↓ Manage a billing dashboard The new one is being built for agents. Our new article explores why x402 is such an important step here. It lets software pay inside the HTTP request itself, which is a much better fit for agent products that need compute, voice, data, and tools on demand. Voice is one of the strongest early use cases. Read more below ↓
dTelecom tweet media
English
34
36
166
4.8K
dTelecom
dTelecom@dtelecom·
No accounts required?.. At first glance, no accounts required sounds like a security risk. Anyone can join a meeting? But the opposite is often true. Accounts create a huge attack surface. Think about: • Stolen credentials • Account takeovers • Phishing logins • Massive user databases for hackers to target Platforms like Zoom and Teams have millions of accounts sitting in central databases. It’s a goldmine for attackers that @dMeetApp removes entirely. Sometimes the safest system is the one that stores less information in the first place. @dMeetApp v.2 available now >> dmeet.org
dTelecom tweet media
English
30
37
123
3.8K
dTelecom
dTelecom@dtelecom·
@trondao let's have your agents get 📢voice and 👂ears x402/MMP ready, own dRTC rails w/ 4× lower latency, and 50-95% cheaper than LiveKit & Deepgram hit us up
English
1
4
39
1.2K
TRON DAO
TRON DAO@trondao·
TRON announced the expansion of its AI Fund from $100 million to $1 billion. The fund will target investments in and acquisitions of early-stage companies building core infrastructure for the agentic economy. The fund will prioritize the development and consolidation of agent identity systems, stablecoin-based payment rails, tokenized real-world assets, and developer tooling for autonomous financial systems. The expansion builds on the fund’s thesis, first outlined in 2023, that AI and blockchain would converge, creating demand for programmable, permissionless financial infrastructure. More details from @Cointelegraph 👇 cointelegraph.com/press-releases…
TRON DAO tweet media
English
90
136
722
174K
dTelecom
dTelecom@dtelecom·
Over 3.3T meeting mins take place on Zoom yearly. Imagine the scale of infra needed to deploy real-time voice translation across those mins. Well, @dMeetApp has a solution: scrap the single giant central pipeline bottleneck. It scales horizontally: • WebRTC sessions are distributed across regional media nodes • STT / translation / TTS run as streaming inference, not batch jobs • stateless workers can be replicated as demand grows • traffic is balanced across regions close to users More sessions → more nodes.
English
33
74
340
8.5K
dTelecom
dTelecom@dtelecom·
We tested swapping ElevenLabs for Kokoro TTS running locally on an Apple Silicon M4 in our voice-agent stack. Results: - 100 ms time to first usable audio - near-zero marginal TTS cost after setup - no network dependency during synthesis That TTS delay is one of the most noticeable parts of the loop. Shaving ms off the speech stage made the interaction feel faster, tighter, and more natural. For interactive voice pipelines, local TTS is now a serious production option. Check details 👇
dTelecom@dtelecom

x.com/i/article/2034…

English
34
51
177
7.1K
dTelecom
dTelecom@dtelecom·
Real-time AI infra. Measured. @dTelecom is now on @Dune • 72M+ minutes processed • 2M+ STT minutes • 8M+ translated characters • 65k+ unique users Live traffic. Live Voice AI (STT/TTS) Real usage by Real Users. Proof ↓ 🔗dune.com/dtelecom_team/…
English
38
42
193
5.5K
dTelecom
dTelecom@dtelecom·
dTelecom is now live on @tempo. AI agents can pay per use for real-time communication - voice calls, video meetings, and Voice AI (STT, TTS). No accounts. No API keys. Powered by @MPP (Machine Payments Protocol).
Tempo@tempo

x.com/i/article/2034…

English
58
70
249
9.2K
dTelecom
dTelecom@dtelecom·
@dwr Congratulations on launching MPP🚀
English
1
4
29
1.1K
Dan Romero
Dan Romero@dwr·
Check out the interactive demo on the MPP docs site mpp dot dev
Dan Romero tweet media
English
12
3
122
9.8K
dTelecom
dTelecom@dtelecom·
We’re now featured in the @CoinbaseDev AgentKit ecosystem. dTelecom is listed among the providers supporting the next generation of AI agents with onchain capabilities. To show what that looks like in practice, we built a Voice Agent example using: - Coinbase AgentKit - a LangChain ReAct agent - dTelecom’s decentralized voice infra Users can speak commands like: “Check my USDC balance” “Send 1 USDC to 0x…” The agent interprets the request, chooses the right onchain tool, executes the action, and speaks the result back. Read more below ↓
dTelecom tweet media
English
8.5K
3.3K
7.8K
39.1K
dTelecom
dTelecom@dtelecom·
Your support bot is deaf, mute, and broke A harsh way to put it, maybe. But for many AI agents, it is true. 1. Deaf: they cannot listen to live audio streams properly 2. Mute: they cannot speak back in a natural real-time session 3. Broke: they still cannot pay for services cleanly without a human account flow somewhere in the stack Sure, they can reason, plan and call tools, but they still cannot participate in conversations. That gap is exactly what interests us at @dTelecom.
English
44
47
277
9.7K