Lahfir (@lahfir) header
Lahfir (@lahfir) profile picture

MTS @vercel labs | built agent-desktop, screen-relay

san diego
Joined April 2026
Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)Photo by Lahfir (@lahfir)
Pinned Tweet
Introducing agent-desktop Computer use for your agents ▲ E2E test desktop apps with agents ▲ Reads apps via the accessibility tree ▲ Drills into part of a window to save tokens ▲ Runs many agents on one app at once ▲ Works on macOS, Windows & Linux npm i -g agent-desktop
82
77
42
1,561
158,643
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
42
54
10
853
125,422
Lahfir retweeted
New emulator: @𝚎𝚖𝚞𝚕𝚊𝚝𝚘𝚛𝚜/𝚊𝚒 Test AI without paying for AI Deterministic but also realistic: streaming, tool calls, latency, errors, varied responses Built for CI, local dev and sandboxes
Introducing local OpenAI & Anthropic APIs in emulate Test AI apps locally with scripted, deterministic replies ➔ OpenAI Responses, Chat & Anthropic Messages ➔ Swap one baseURL, keep your real SDK ➔ Tool loops, structured output & reasoning ➔ Inject 429s, stream timing, stalls & cuts ➔ Seeded variants repeat the same order every run ➔ No API keys, no network, zero token spend npx emulate --service ai
48
29
5
732
28,815
Introducing local OpenAI & Anthropic APIs in emulate Test AI apps locally with scripted, deterministic replies ➔ OpenAI Responses, Chat & Anthropic Messages ➔ Swap one baseURL, keep your real SDK ➔ Tool loops, structured output & reasoning ➔ Inject 429s, stream timing, stalls & cuts ➔ Seeded variants repeat the same order every run ➔ No API keys, no network, zero token spend npx emulate --service ai
17
18
11
216
35,953
Lahfir retweeted
Agents can drive the desktop too screen-relay screenshot shot.png screen-relay type "hello" screen-relay key super+Left screen-relay key ctrl+alt+t ▲ Skills ship with the CLI: `screen-relay skills get core`
2
1
38
6,982
Introducing screen-relay Stream and control any desktop from your browser ➔ Up to 120 fps ➔ 1080p and higher ➔ Sends only what changed ➔ Adapts to your connection ➔ Agents drive it from a CLI, SDK or API ➔ Works on macOS, Windows & Linux npm i -g screen-relay
68
95
21
1,537
137,983
Lahfir retweeted
Introducing eve 1.0 Build production agents.
eve is generally available with v1.0. • Build agents on your own harness • Ship multi-tenant, multiplayer agents • Govern & trace sessions with Agent Runs • Render agent UIs via Next.js, Nuxt & more Read the blog → vercel.com/blog/eve-is-now-g…
46
62
13
1,091
144,293
Replying to @ktiays
awesome!! uikit.even.how
1
3
39
6,196
portless v0.18 now has JSON mode. Add --json to list, get or doctor and scripts can read your routes without parsing a table. Also: → mistyped flags are caught before anything runs → stdout holds only your app's output → clean asks before execution
1
3
1
22
1,625
Lahfir retweeted
Say hi to the new agent-browser.dev 👋 Featuring agent-browser + agent-desktop Designed by @jamesvclements
19
31
1
831
40,559