Agent Browser is now live on the Telnyx API.
Give your AI agent its own browser.
One API call starts an isolated, short-lived Chromium session that can only visit the sites you allow.
Your agent can open pages, read them, click, fill in forms and take screenshots, so it can that have no API.
The boundaries are set before the agent starts.
You create a session and list the domains it may visit.
Any navigation outside the list fails with policy_denied.
The agent reads each page as a text tree of labeled elements, then clicks, fills and selects by element reference.
Every action returns structured JSON, ready to hand back to a model.
A batch request runs up to 50 actions in order.
Sessions last 15 minutes by default and close after 5 idle minutes, so nothing carries over between tasks.
Read more about our Agent Browser here:
Docs: developers.telnyx.com/docs/agent-bro…
Release note: telnyx.com/release-notes/…
Give your AI agent its own browser.
One API call starts an isolated, short-lived Chromium session that can only visit the sites you allow.
Your agent can open pages, read them, click, fill in forms and take screenshots, so it can that have no API.
The boundaries are set before the agent starts.
You create a session and list the domains it may visit.
Any navigation outside the list fails with policy_denied.
The agent reads each page as a text tree of labeled elements, then clicks, fills and selects by element reference.
Every action returns structured JSON, ready to hand back to a model.
A batch request runs up to 50 actions in order.
Sessions last 15 minutes by default and close after 5 idle minutes, so nothing carries over between tasks.
Read more about our Agent Browser here:
Docs: developers.telnyx.com/docs/agent-bro…
Release note: telnyx.com/release-notes/…
0
2
6
538
If the infrastructure is rented, you don’t have full control over the third-party vendor’s infrastructure.
If the answer includes another continent, your compliance team has a problem to explain.
If it is "regional hosting is coming", you are buying a promise.
Telnyx runs the GPUs in Dubai. Data is processed in the UAE. Not an add-on. It is how the platform is built.
If your board has already asked where your AI data is processed, contact us. We will give you a clear answer.
Visit:
#𝗧𝗲𝗹𝗻𝘆𝘅 #𝗠𝗘𝗡𝗔 #𝗠𝗶𝗱𝗱𝗹𝗲𝗘𝗮𝘀𝘁 #𝗔𝗜 #𝗔𝗜𝗔𝗴𝗲𝗻𝘁 #𝗔𝗿𝗮𝗯𝗶𝗰𝗔𝗜 #𝗩𝗼𝗶𝗰𝗲𝗔𝗜 #𝗗𝗮𝘁𝗮𝗦𝗼𝘃𝗲𝗿𝗲𝗶𝗴𝗻𝘁𝘆
If the answer includes another continent, your compliance team has a problem to explain.
If it is "regional hosting is coming", you are buying a promise.
Telnyx runs the GPUs in Dubai. Data is processed in the UAE. Not an add-on. It is how the platform is built.
If your board has already asked where your AI data is processed, contact us. We will give you a clear answer.
Visit:
#𝗧𝗲𝗹𝗻𝘆𝘅 #𝗠𝗘𝗡𝗔 #𝗠𝗶𝗱𝗱𝗹𝗲𝗘𝗮𝘀𝘁 #𝗔𝗜 #𝗔𝗜𝗔𝗴𝗲𝗻𝘁 #𝗔𝗿𝗮𝗯𝗶𝗰𝗔𝗜 #𝗩𝗼𝗶𝗰𝗲𝗔𝗜 #𝗗𝗮𝘁𝗮𝗦𝗼𝘃𝗲𝗿𝗲𝗶𝗴𝗻𝘁𝘆
0
2
6
224
Press 1 for billing. Press 4 for account changes. Press 9 for more options. Then hold.
Every caller knows the phone menu maze. Layer after layer, transfer after transfer, and still the wrong team.
A voice AI agent skips all of it.
Say what you need, and it routes you to the ime.
#VoiceAI #CustomerExperience #AIAgents #Telecom #Telnyx
Every caller knows the phone menu maze. Layer after layer, transfer after transfer, and still the wrong team.
A voice AI agent skips all of it.
Say what you need, and it routes you to the ime.
#VoiceAI #CustomerExperience #AIAgents #Telecom #Telnyx
1
0
4
333
𝗪𝗵𝗮𝘁 𝗶𝗳 𝘆𝗼𝘂𝗿 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁 𝗰𝗼𝘂𝗹𝗱 𝗿𝗲𝗺𝗲𝗺𝗯𝗲𝗿 𝘄𝗵𝗮𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀, 𝗲𝘃𝗲𝗻 𝗮𝗳𝘁𝗲𝗿 𝗮 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻 𝗲𝗻𝗱𝘀?
With Telnyx Agent Memory, you can build AI agents that retain useful information across sessions and recall relevant facts when Python walkthrough on YouTube to see how it works 👇
dub.sh/ewhqBn0
#AI #AIagents #Telnyx #Python
With Telnyx Agent Memory, you can build AI agents that retain useful information across sessions and recall relevant facts when Python walkthrough on YouTube to see how it works 👇
dub.sh/ewhqBn0
#AI #AIagents #Telnyx #Python
0
1
5
368
Meta's WhatsApp rates vary by country. Telnyx's fee doesn't.
Just $0.004/message on top of Meta's rate. No monthly platform fee or minimum spend.
Calculate your costs → telnyx.com/pricing/whatsa…
#WhatsAppAPI #CPaaS
Just $0.004/message on top of Meta's rate. No monthly platform fee or minimum spend.
Calculate your costs → telnyx.com/pricing/whatsa…
#WhatsAppAPI #CPaaS
0
0
3
260
Wondering what to build with the new Telnyx Meeting API?
Here are 20 ideas.
They cover recruiting, sales, support, research, and education.
Some assistants talk in the meeting, some work during the call, and some write up what happened afterward.
The Meeting Bot API is in at
developers.telnyx.com/docs/meeting
What are you building with Telnyx meeting API?
Here are 20 ideas.
They cover recruiting, sales, support, research, and education.
Some assistants talk in the meeting, some work during the call, and some write up what happened afterward.
The Meeting Bot API is in at
developers.telnyx.com/docs/meeting
What are you building with Telnyx meeting API?
1
1
6
547
𝗬𝗼𝘂𝗿 𝗔𝗜 𝗯𝗶𝗹𝗹 𝗶𝘀 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗴𝗿𝗼𝘄𝗶𝗻𝗴. 𝗗𝗼 𝘆𝗼𝘂 𝗸𝗻𝗼𝘄 𝘄𝗵𝗮𝘁’𝘀 𝗱𝗿𝗶𝘃𝗶𝗻𝗴 𝗶𝘁?
Voice Connect Sydney is this week. We’ll unpack the biggest drivers of AI spend, with real pricing on screen and practical ways to bring costs under control.
You’ll ice AI stories from our partners before our team takes you through the token spend playbook.
Wednesday 21 October | 12:00PM to 1:30PM | ACS Harbour City Labs, Barangaroo
➡️ Last spots. Register here: tlyx.co/1bk
#AIEngineering #Sydney #VoiceAIConnect
Voice Connect Sydney is this week. We’ll unpack the biggest drivers of AI spend, with real pricing on screen and practical ways to bring costs under control.
You’ll ice AI stories from our partners before our team takes you through the token spend playbook.
Wednesday 21 October | 12:00PM to 1:30PM | ACS Harbour City Labs, Barangaroo
➡️ Last spots. Register here: tlyx.co/1bk
#AIEngineering #Sydney #VoiceAIConnect
1
0
5
330
Your data, processed in the #𝗠𝗶𝗱𝗱𝗹𝗲𝗘𝗮𝘀𝘁
In-house AI teams have two fixed costs: people and tokens.
You control the first. The second is set by whoever hosts your model, and it rises with every feature you ship.
Open-source models change that. You can inspect them, own data and run them where you choose. No dependence on one vendor's price list or one vendor's roadmap.
Telnyx runs the top-notch open-source models like Kimi, GLM, and DeepSeek on owned GPUs in #𝗗𝘂𝗯𝗮𝗶, from $0.13 per million tokens, with data processed in the region.
One API. Your code stays yours.
Contact us and we will provide you with a cost comparison for your current workload:
#𝗧𝗲𝗹𝗻𝘆𝘅 #𝗜𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 #𝗢𝗽𝗲𝗻𝗦𝗼𝘂𝗿𝗰𝗲𝗠𝗼𝗱𝗲𝗹𝘀 #𝗔𝗜 #𝗟𝗟𝗠 #𝗖𝗹𝗼𝘀𝗲𝗱𝗠𝗼𝗱𝗲𝗹𝘀 #𝗠𝗘𝗡𝗔 #𝗚𝗖𝗖 #𝗨𝗔𝗘 #𝗦𝗮𝘂𝗱𝗶
In-house AI teams have two fixed costs: people and tokens.
You control the first. The second is set by whoever hosts your model, and it rises with every feature you ship.
Open-source models change that. You can inspect them, own data and run them where you choose. No dependence on one vendor's price list or one vendor's roadmap.
Telnyx runs the top-notch open-source models like Kimi, GLM, and DeepSeek on owned GPUs in #𝗗𝘂𝗯𝗮𝗶, from $0.13 per million tokens, with data processed in the region.
One API. Your code stays yours.
Contact us and we will provide you with a cost comparison for your current workload:
#𝗧𝗲𝗹𝗻𝘆𝘅 #𝗜𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 #𝗢𝗽𝗲𝗻𝗦𝗼𝘂𝗿𝗰𝗲𝗠𝗼𝗱𝗲𝗹𝘀 #𝗔𝗜 #𝗟𝗟𝗠 #𝗖𝗹𝗼𝘀𝗲𝗱𝗠𝗼𝗱𝗲𝗹𝘀 #𝗠𝗘𝗡𝗔 #𝗚𝗖𝗖 #𝗨𝗔𝗘 #𝗦𝗮𝘂𝗱𝗶
2
1
5
642
India's voice AI stack is coming together.
With Programmable Voice now available in Mumbai, developers can bring together local telephony, Indian mobile numbers, speech models, and AI inference on the Telnyx platform.
Whether you're building voice AI agents for Indian consumers re applications, the building blocks are now in place to run voice experiences for India on one network.
One platform. From the phone number to the AI.
Release Note: dub.sh/MocEJgr
#VoiceAI #AIinIndia #ProgrammableVoice
With Programmable Voice now available in Mumbai, developers can bring together local telephony, Indian mobile numbers, speech models, and AI inference on the Telnyx platform.
Whether you're building voice AI agents for Indian consumers re applications, the building blocks are now in place to run voice experiences for India on one network.
One platform. From the phone number to the AI.
Release Note: dub.sh/MocEJgr
#VoiceAI #AIinIndia #ProgrammableVoice
1
1
7
378
Same model ID, GLM-5.3 Flash, on Telnyx and Modal Shared.
240 requests each.
p50 end-to-end: 4.92s vs 8.75s
p95 end-to-end: 11.91s vs 20.62s
Output speed: 138.1 vs 66.2 tok/s
Benching by @Coffee_and_NLP. Method 👉 telnyx.com/resources/glm5…
#Inference #OpenSource
240 requests each.
p50 end-to-end: 4.92s vs 8.75s
p95 end-to-end: 11.91s vs 20.62s
Output speed: 138.1 vs 66.2 tok/s
Benching by @Coffee_and_NLP. Method 👉 telnyx.com/resources/glm5…
#Inference #OpenSource
0
1
2
406
If you’ve ever built a voice agent, you’ve met this monster.
Five vendors. Five invoices. Five places to look when a call sounds wrong at 2 a.m.
We drew it so you can send it to whoever signed off on the architecture: telnyx.com/landing/franke…
#VoiceAI #AIAgents #Frankenstack #DevEx #RealTimeAI
Five vendors. Five invoices. Five places to look when a call sounds wrong at 2 a.m.
We drew it so you can send it to whoever signed off on the architecture: telnyx.com/landing/franke…
#VoiceAI #AIAgents #Frankenstack #DevEx #RealTimeAI
0
2
4
392