The Teneo CLI installs from wherever you already work. There are now setup guides for seven of them.
AI coding agents:
- Claude Code: https://t.co/0CXaQdNc9N
- Codex: https://t.co/K9vZ8wuyt1
- Manus: https://t.co/JVCAyP38A6
IDEs:
- Cursor: https://t.co/JWOAmPEHwE
- Antigravity: https://t.co/e8vMwuuLUg
-VS Code: https://t.co/7xnxHXDZHQ
Agent platforms
Hermes: https://t.co/g9A5HG89S3
Each guide walks through installing the CLI, then listing and querying 50+ Teneo agents in plain English: scrape social data, swap and bridge tokens, pull on-chain data. Some agents are free; paid queries settle per call in USDC via x402.
Two more are in progress:
OpenClaw and Teneo Father (query agents straight from Telegram).
Full list: https://t.co/bnbIJYAAVk
What are we missing? If you'd run the CLI in a tool that isn't here, tell us.
Agentic AI Is Getting Counterparties, Not Just Rails
- Stripe closed a ~$7.5B acquisition of OpenRouter, the router moving 10T+ tokens a day across 400+ models.
- TAO went natively live on Base via Chainlink CCIP, no admin key, so on-chain agents can now hold and deploy it directly.
- NVIDIA is reportedly in talks to invest several hundred million in Cloverleaf, a powered-land developer with 7GW+ already sold.
Three stories, one shift: last quarter the agent economy standardised its payment rails. This week it's standardising who's on the other side of the trade.
Occupancy is now the story: agents as paying customers, agents as asset holders, capital chasing the power those agents will run on.
On Teneo, hundreds of live agents already settle exactly this way: named commands, quoted per call, paid in USDC through x402.
The rails are commodity now.
Occupying them is the game!
Read the full round-up here: 👇
https://t.co/uv59hsAySN
The Teneo Beacon now runs headless on your Terminal!
Turn the Raspberry Pi, home server or old laptop you already own into an always-on Teneo Beacon.
Install with one command. Run teneo-beacon and scan the QR code while signed in to the Hub. Install the service, and it starts on boot and reconnects after an outage.
One limit matters: earning is one slot per internet connection, not per device. Two Beacons on the same home network do not stack. The CLI lets a low-power machine hold that slot around the clock.
Copy the install command from the official page, not a screenshot or reply. It runs with shell access, and the source link sits beside it.
Full details are here👉: https://t.co/SUaIs95h9a
Inference isn’t the scarce resource anymore.
In the last five weeks alone, the cost of running an agent dropped four times. But giving that agent a real, reliable view of the live web? That didn’t get any cheaper.
A local model can plan and call tools all day for pennies. But it’s still stuck behind one home connection — no rotation, no geographic coverage, no way to confirm the page it just pulled is what a real person in that country actually sees.
Weights can be copied overnight. Live web access can’t.
That’s the gap Teneo Beacons fill. Hundreds of agents are already running production queries through them, settling every call in USDC.
The network is live. The market is forming on top of it.
Keep your Beacon running.
Agentic AI Is Leaving the chat window!
- MetaMask gave AI agents a self-custodial wallet with spending limits.
- A gym-booking agent found an exposed API and removed another person from a waitlist.
- Riot Platforms signed a 20-year lease for 191 MW of capacity with a leading frontier AI lab.
Three stories, one shift: agents are leaving the chat window and entering systems where actions have consequences.
Permissions are now infrastructure: spend limits, scoped tools, paid data and enough records to reconstruct an action.
On Teneo, data agents expose named commands, quote each query and settle in USDC through x402.
CoinMarketCap Agent `quote`: $0.005 base, with the signed total returned before settlement.
Capability is moving fast. Accountability has to become part of the transaction path.
Read the full post here: 👇
https://t.co/REEjbnZU7R
Staking asks you to lock it up and wait. Your capital goes away for ninety days and the yield is whatever it is when it comes back.
Your Teneo Beacon locks up nothing. 10 Fragments per unique IP per hour, from bandwidth you're already paying for.
The only thing that compounds is showing up. Claim and boost every 8 hours and Beacon Power climbs towards 3.00x. Miss the windows and it decays back down.
No lockup. Just don't go quiet.
Keep your Beacon running!
Download Beacon: https://t.co/YY4D4tiE8r
Monday reminder, mostly for people who joined in the last few weeks: the updated 2026 roadmap is live, and the first half of it is already done.
The network runs in production today. Community nodes supply public web data across multiple platforms, hundreds of agents are deployed and every paid query settles in USDC over x402.
None of that is a future milestone!
What the rest of the year does is turn that network into a market. Enterprise customers are onboarding now!
The operator side of that does not change: demand gets sold on coverage and uptime, and both come from nodes staying online. Keep your Beacon running.
Full roadmap: https://t.co/sPRldc7tkc
The cost of running an agent fell four separate times in five weeks.
- 30 July: OpenAI cut GPT-5.6 Luna by 80%, to $0.20 per million input tokens
- 31 July: DeepSeek's V4-Flash beta beat its own larger V4-Pro preview on agent work, same size, redone post-training
- 3 August: Alibaba priced Qwen3.8-Max, 2.4T parameters, at $2 input, and said the weights are going open
- 4 August: Liquid AI shipped a 2.6B agentic model that plans and calls tools on a phone at 30 tokens a second, no marginal cost
None of the four made the web any easier to read. A local agent can think all day for free and still has one home connection, no rotation, and no way to know whether it was served the page a real person in that country would see.
Weights can be copied in an afternoon. Access to the live web cannot.
The full post has the pricing on all four and what follows from them: free reasoning, and a tightening supply of anything true to reason about.
https://t.co/HUuA9umnb8
Agents are impressive. Multi-agent systems are where things get interesting.
As AI systems evolve from single agents into networks of specialists, a new question follows:
How do agents decide who does what?
That’s a problem we’ve been exploring with Symphony.
Symphony brought multi-agent collaboration across heterogeneous environments. To take that further, we released Symphony-Coord, where coordination and specialization adapt through task context, performance, and feedback.
https://t.co/7qdcmO40RU
Teneo is on Looplay's AMA #7 today. 3pm UTC, X Spaces.
The topic is what actually breaks when vibe-coded software reaches production.
Our Head of Marketing @Sarut0biSasuke joins @Hertzflow_xyz and @zambodotdev to work through it.
We run hundreds of agents in production. The generated code is rarely the failure. The failure is the contract between components: what one agent assumed about another's output, unwritten and untested, holding fine right up until the input shape changes.
Hallucinations are the visible problem. Undocumented assumptions are the expensive one.
Hosted by @looplaygg_ .
Set a reminder!
Most AI data tools make you move into their stack before you find out if the data is any good.
Teneo's CLI speaks MCP. One command, into whatever you're already running.
Claude Code. Cursor. A terminal. A self-hosted gateway like OpenClaw. A standalone runtime like Hermes.
Behind it, hundreds of live agents. Market data, on-chain analytics, public X data in real time.
Pay per call in USDC over x402. No API key, no subscription, no seat.
Nothing to migrate!
Check out our getting started guides here: 👇
https://t.co/bnbIJYAAVk
DePIN passed the oracle sector in market cap this year.
The category holds 264 tokens against the oracle sector's roughly 60, across more than 650 live projects.
Most people still stall at three names, because DePIN gets used as if it labels one thing. It labels five, with almost nothing in common except the payment model:
- Compute: GPU cycles from idle machines
- Storage: disk space that stays retrievable
- Wireless: coverage: from hotspots in homes and shops
- Sensors and mapping: dashcams, weather, air quality
- Data infrastructure: residential connections reaching the live web from a real place
Four of those five can be scaled with capital. Buy the GPUs, buy the drives, buy the hotspots.
The fifth has to be assembled one household at a time. That is the column Teneo Beacon is in.
The full post covers why that column changed in 2026, and what a Beacon node actually supplies.
Your Beacon is already in this category.
Read the full post here 👇:
https://t.co/tEqwS3Cg45
Airdrop farming is a part-time job. Quest timers, eligibility checks, campaign windows you have to be awake for.
Your Teneo Beacon earns 10 Fragments per unique IP per hour while you do nothing at all.
Claim and boost every 8 hours and your Beacon Power
multiplier, 1.00x to 3.00x, applies to everything you have banked since the last cycle.
Consistency is the only input.
Keep your Beacon running!
Download Beacon: https://t.co/YY4D4tiE8r
What Teneo shipped in July, with the numbers:
- Season 2 closed with 706 verified winners.
- Season 3 added per-node verification tasks to block bots
- Season 3 opened for 60 days on a cleaner leaderboard.
- Beacon v0.4.7 patched Android crashes that ended sessions early.
- $1,000 USDC was reserved for Beacon runners on verified Indonesian IPs, weighted by uptime.
- Five new posts covered Beacon operations, the DexScreener agent, and a per-query meeting-prep workflow.
- CoinMarketCap featured Teneo through CMC Labs for the third time this year.
Season 3 is live!
Keep your Beacon running!
Read the full July recap:
https://t.co/XDbPVrZthS
The Week the Agent Economy Got Real
Bot traffic passed human traffic on the open web for the first time ever.
Coinbase shipped agent payments to every business on its platform.
Every major wallet went agent-native inside four weeks.
None of it solved what agents are actually short of.
The last eight weeks:
3 June: Cloudflare's CEO reports automated traffic at 57.5% of HTTP requests against 42.5% human, roughly 18 months ahead of his own forecast
23 July: Coinbase opens agent-initiated USDC acceptance to every Business merchant, plus an x402 SDK that adds payment acceptance in three lines of code
Same window: MetaMask, OKX, Coinbase and BNB Chain all go agent-native, and 27 companies back a dispute court for agents
Wallets, rails, chains, arbitration. All standardised, in public, on open specs.
Not one of them lets an agent read a page that doesn't want to be read.
And on 15 September, Cloudflare starts blocking mixed-use AI crawlers by default.
That's the constraint we've been building against:
Full read:
https://t.co/IwgRnSUq8c
What if your Beacon bandwidth paid for your VPN?
Right now, Beacon nodes share unused residential bandwidth with the network.
That bandwidth already looks a lot like what a VPN needs: real connections, in real places, all over the world.
So we're asking ourselves: instead of paying for a VPN with a monthly bill, could you pay for it with the bandwidth you're already contributing?
Still very early. Nothing to download, no launch date.
We wanted to share it while we're building it, not after.
Full writeup: https://t.co/HdI0OTXacf
Season Update: Indonesia Pool 🇮🇩
$1,000 USDC from this season's reward pool is reserved exclusively for Beacon runners on Indonesian IPs.
Your share is based on uptime, so the more consistently your Beacon runs, the bigger your cut.
Eligibility is tied to genuine connections in Indonesia.
Full mechanics drop this week.
If you're in Indonesia, get your Beacon running now.
Need to download the Beacon?
Get it running in 2 minutes: https://t.co/YY4D4tiE8r
Live AMA with OpenX Network tomorrow, Tuesday July 28 at 15:00 UTC.
Our Head of Marketing @Sarut0biSasuke joins @AshtonH to talk decentralized AI and DePIN from the operator side: what it takes to run a network with hundreds of agents live in production, community nodes supplying the data, and every query settled in USDC over x402.
Bring questions about running a Beacon, building on the Go SDK, or where the agent economy goes next.
We'll take them live.
Most of our blog posts show what ONE Teneo agent does.
This one shows what happens when you chain TWO of them into an actual workflow.
Our Head of Marketing @Sarut0biSasuke runs meeting prep using Hermes from @NousResearch
- Every 10 minutes it checks Google calendar
- For any attendee it doesn't know, it calls the Teneo LinkedIn Agent (enrich_url <url>) for who they are
- Then the Teneo X Agent to find their project's X account
- Then it drops a one-page brief before the meeting starts.
The cost:
LinkedIn enrichment costs $0.006/query, X context from $0.002/query, both settled in USDC over x402 on Base.
For ten meetings a week, three attendees each, it will set you back about a dollar a month.
Sales Navigator is $100+/month and still makes you do the reading yourself.
The full post covers the workflow end to end:
https://t.co/jroR40XxCJ
🎉 The Teneo Creator Challenge has officially come to an end!
A huge thank you to everyone who took the time to create and share videos. It was great seeing so many creative ways to explain the Beacon and introduce new people to Teneo.
After carefully reviewing all valid entries, we've selected our 13 winners. The selection wasn't based on views alone, but on the overall quality, clarity, engagement, and how well each video could inspire someone to start running a Beacon.
If you're one of the winners, we'll be reaching out shortly through the social media account you used to participate or via the Discord username you included in your submission to collect your wallet details.
Congratulations to all the winners, and thank you again to everyone who participated. 💚 #KeepYourBeaconRunning