ry retweeted
Shopify's CEO gives his AI agents their own computer.
Tobi Lütke just shared wrapvm: a full desktop VM built for agents.
• A real Linux desktop your agent can see and control
• Its own Chrome, already wired up for browser automation
• Computer-use tools so it can click through any app
• One machine per project folder, so agents never touch your real computer
• You watch everything live from your browser, and sound can be recorded too
Your agent gets a sandbox to break things in. Your laptop stays yours.
github.com/tobi/wrapvm
1
1
35
3,201
ry retweeted
Start Tembo sessions from your own code.
The Tembo API, TypeScript SDK and MCP server let a script, an internal tool or another agent hand work to Tembo, then follow the session and its diff.
Read the announcement → tembo.io/blog/announcin…
The Tembo API, TypeScript SDK and MCP server let a script, an internal tool or another agent hand work to Tembo, then follow the session and its diff.
Read the announcement → tembo.io/blog/announcin…
1
3
20
2,324
ry retweeted
it's wild how much better voxl is than VNC/screen sharing. This has become my default remote viewer. also built nice little ipad/fold8 clients for it.
github.com/ericmigi/voxl-…
github.com/ericmigi/voxl-…
github.com/ericmigi/voxl-…
github.com/ericmigi/voxl-…
Introducing voxl
extremely fast remote desktop streaming written in rust
→ 60 fps + low latency
→ macOS, Windows and Linux
→ One WebSocket, so it works through proxies and sandbox portals
→ Agent APIs: /screenshot, /record, /wait
10
19
479
20,253
skills 📈
New in Tembo ✨
• A new skill editor with a file tree, scripts, and folder import
• Google Docs, Sheets, and Drive integrations
• See when your team is most active with Insights active hours
• Starting sessions from slack is now ~7x faster
Read the full changelog ↓
tembo.io/changelog#nov-…
0
0
5
404
We've worked long and hard to try to get startup times as low as possible, even when you're loaded up with Repos, MCPs, and Skills :)
The world's fastest cloud coding agent.
From Enter to a running app:
2.20s → Tembo
26.88s → Devin
61.25s → Cursor Cloud Agents
The VM is booted and the repo is set up before you even ask.
How it works ↓
tembo.io/blog/agent-env…
2
0
10
679
ry retweeted
cursor went to spacex - ona went to openai - conductor just went to meta
every time a startup in our space gets acquired people ask if that's the plan
it's not really what i think about
this is already all i wanted
- freedom to work from anywhere
- owning what we build
- it: design, code, gtm
- having fun with people i care about
every time a startup in our space gets acquired people ask if that's the plan
it's not really what i think about
this is already all i wanted
- freedom to work from anywhere
- owning what we build
- it: design, code, gtm
- having fun with people i care about
11
3
71
4,394
see it live in @tembo - self-serve free trial is back
and if you DM any of us feedback on the product, we'll grant you bonus credits
and if you DM any of us feedback on the product, we'll grant you bonus credits
we just hit 1k stars on voxl!
thanks for all the support on this post, glad everyone likes it
come give the fastest desktop streaming for software factories a try: github.com/tembo/voxl
1
0
6
360
ry retweeted
congrats charlie and team
if you need a new home, @tembo:
- self-host on your own infra
- runs claude code, codex, cursor, opencode, amp, pi, grok build, antigravity
- run as many agents at once as you want
- each gets its own linux VM
- no git worktrees
- any model: claude, gpt, gemini, grok, open source
- use your claude, chatgpt or supergrok plan, or API keys
- review diffs before the PR
- sessions keep running when you close your laptop
- preview builds in your browser
- 60 FPS remote desktop
- fork a teammate's session
- agent templates
- API, SDK and MCP
- agents have their own accounts
- github, jira, linear and more
- VMs auto scale up to 128 GB mem
- VMs ~20% cheaper than E2B or Daytona
tembo.io
if you need a new home, @tembo:
- self-host on your own infra
- runs claude code, codex, cursor, opencode, amp, pi, grok build, antigravity
- run as many agents at once as you want
- each gets its own linux VM
- no git worktrees
- any model: claude, gpt, gemini, grok, open source
- use your claude, chatgpt or supergrok plan, or API keys
- review diffs before the PR
- sessions keep running when you close your laptop
- preview builds in your browser
- 60 FPS remote desktop
- fork a teammate's session
- agent templates
- API, SDK and MCP
- agents have their own accounts
- github, jira, linear and more
- VMs auto scale up to 128 GB mem
- VMs ~20% cheaper than E2B or Daytona
tembo.io
I am very excited to share that Conductor is joining Meta!
We have some big, crazy plans for the next interface. If you liked Conductor, you’re gonna love what’s coming.
We’re super excited to work with @natfriedman and @alexandr_wang at MSL to build it. They’ve been on an convinced MSL has the highest talent density of all the labs.
A huge thank you to our earliest believers @aaron_epstein, @ilyasu, and @nabeel. And to everyone here who tried Conductor, sent us feedback, and prompt requests - you made this so much fun. Thank you ❤️🚂
Choo choo!
8
5
26
2,119
ry retweeted
we were harness agnostic all the way back in august of 2025
being able to just switch to the best model + harness combo whenever you want is table stakes now
being able to just switch to the best model + harness combo whenever you want is table stakes now
Stop making your team pick one coding agent.
Tembo runs them in the cloud, on whichever model you want:
Claude Code
Codex
Cursor
OpenCode
Amp
Pi
+ more
Already paying for Claude, ChatGPT or SuperGrok? Plug that plan in.
How it works → tembo.io/blog/run-any-c…
0
1
4
425
no reason to have choppy cloud browsers anymore
voxl v0.3.3 just released with some small improvements to our macos support!
here is voxl running in @tembo streaming my personal macbook live to a actively running tembo session
(yes i can play minecraft in 60fps while my agents work)
1
0
12
455
ry retweeted
forgot to mention insanely fast starts
our team has been merging PRs 3x faster these last few weeks bc of @tembo
our team has been merging PRs 3x faster these last few weeks bc of @tembo
Introducing voxl
extremely fast remote desktop streaming written in rust
→ 60 fps + low latency
→ macOS, Windows and Linux
→ One WebSocket, so it works through proxies and sandbox portals
→ Agent APIs: /screenshot, /record, /wait
1
5
11
917
ry retweeted
this is huge for frontend
animations and moving pieces are half the job - over normal VNC streaming they come through choppy
voxl, our in-house stream engine, runs at a near-constant 60 FPS
you will FEEL the difference
want tembo for your org? dm me, i'll get you sorted
animations and moving pieces are half the job - over normal VNC streaming they come through choppy
voxl, our in-house stream engine, runs at a near-constant 60 FPS
you will FEEL the difference
want tembo for your org? dm me, i'll get you sorted
Introducing voxl
extremely fast remote desktop streaming written in rust
→ 60 fps + low latency
→ macOS, Windows and Linux
→ One WebSocket, so it works through proxies and sandbox portals
→ Agent APIs: /screenshot, /record, /wait
0
1
20
2,548
ry retweeted
See what your agent changed before there's a PR.
A Tembo session's Changes tab shows the diff. Comment on any line, send your review to the agent, and open the PR when it's ready.
How it works → tembo.io/blog/review-co…
A Tembo session's Changes tab shows the diff. Comment on any line, send your review to the agent, and open the PR when it's ready.
How it works → tembo.io/blog/review-co…
2
4
10
2,340
Recently did a PR analysis for a customer; 9 of 10 PRs were fixes/defense. Many many teams are in catch up mode, not value creation mode.
It's not just engineering either. A CX team manually walking customers through half-built onboarding is playing defense.
Missing out on huge opportunity to attack their market.
It's not just engineering either. A CX team manually walking customers through half-built onboarding is playing defense.
Missing out on huge opportunity to attack their market.
5
0
4
505