Conductor, a Mac app for running several AI coding agents side by side, said on November 4 that it is joining Meta. Founder Charlie Holtz posted the news on X at about 10:18 AM PT, and Conductor's announcement page lists a shutdown schedule that starts the same day. By about 3:20 AM PT on November 5, Holtz's X post showed 4,067 likes, 726 replies and 662,312 views, and the Hacker News thread had 181 points and 199 comments. We covered the deal itself in Meta Buys Conductor; Cloud Shuts Nov 20. This roundup is about the reaction, and it is mostly about what happens to the people who built their workday on the product. Reddit cannot be read from our tools, so it is not covered.
What Conductor said
The facts below come from Conductor's announcement and FAQ. New sign-ups and downloads closed on November 4. Pro users can keep using pro features for free, and active subscribers get a refund. Conductor Cloud goes read-only on November 13 and shuts down on November 20, when deletion of Cloud user data begins. Local Conductor keeps getting new models and critical bug fixes until early 2027, and then starts winding down. Conductor says it built a "Transfer to local" option for cloud workspaces, and that it has "never sold or trained on Conductor user data."
Holtz's post on X said the team will work with Nat Friedman and Alexandr Wang on what he called "the next interface." Alexandr Wang replied at about 6:38 PM PT with "welcome conductor team to meta", and that post showed 516 likes and about 44,000 views at the time we checked.
Hacker News: the shutdown dates hit hardest
Most of the thread is from users, not onlookers. BowBun said Conductor was a main development tool and asked why it could not be open sourced. stevenou wrote that "Having cloud webhooks + workspace/session API + isolated VMs has been a massive productivity boost for me," and asked for something that matches Conductor Cloud, "not just Conductor Local." iamunr said their team had gone "all in" on the cloud setup the day before, and now was "doing a refund + moving the org to t3.codes instead."
Several people questioned the speed of the shutdown. ayayronf called the timeline to migrate off Conductor Cloud "very aggressive." Raed667 said the shutdown timeline "seems rather rushed." markerz quoted the dates from the announcement and wrote "Flashback to Cloudflare acquiring Deno and hiding the shutdown at the bottom of their announcement." nickfmatthews reported a gap in the other direction: sign-ups are closed, and they could not find a way to try the Mac app locally without an account.
Some replies were about trust in the founders. dayone1 asked "why sellout your company and your customers and community?", and doublerabbit asked Holtz directly why the company sold. Holtz, who joined the thread as Charlieholtz, thanked the community and linked the July 2025 Hacker News post that launched Conductor. jahala said they refuse to use Meta products and might have Claude reverse engineer Conductor over the next month, while also congratulating Holtz in a later reply. BowBun wrote later that their team had by the afternoon forked and reverse engineered the whole project using LLMs. That is the commenter's own account, and we have not seen the code.
Hacker News: where people are going
The most common question was what to switch to. paultannenbaum opened the longest sub-thread with it, and the answers named T3 Code, Orca, Herdr, Paseo, boxes.dev and several others. 2001zhaozhao wrote that T3 Code is "by far the best one I have tried that isn't glued to a specific model provider," and added that they are building their own tool in the same area. nwparker said they work on Orca, which can import from Conductor. crystaln said they had moved to Paseo "long ago" because of its mobile app and open source code, and that Conductor "began to feel stale, pricey."
mattmaroon noticed the pattern: "The number of comments in here saying 'founder of xyz here, if you're looking for a replacement we make a similar product' is illustrative." markush_ went further and read the number of founder replies as a sign that "there is clearly no moat in this." Readers should treat the recommendations in this thread the same way: several come from people who sell or build the tool they name, and a few say so. The same applies to the X replies below.
X: congratulations, then the migration question
On X the first reaction came from users. AsherDale13, at about 11:24 AM PT, wrote that they "run basically all my agents through it" and asked for suggestions. pulkit_gupta2 wrote at about 7:20 PM PT that "when they launched cloud we went from ~5 prs/day to 90+ prs/day," that alternatives "almost all" fall short, especially on UI and UX, and listed what they need next: multiplayer with API access, agent to agent communication and support for a pool of Claude and Codex subscriptions. That post showed 45 likes and 48 replies. This is one team's account of its own output, and we did not check it.
valerioHQ called it "devastating news (for the users)" and "by far the best ADE I've ever tried," adding "I wish there was a smoother transition for users." p44v9n, who describes a role as a product designer, said Conductor let them run multiple terminals with worktree setup scripts and work from a combined pool of token limits, called it "gamechanging", and wrote "well done to the Meta team for a big nab."
Others used it as a lesson. emaann28 wrote that "Building a useful product doesn't guarantee that the product itself will survive." TbYoutuber suggested a new question for the AI tool checklist: "If this startup gets acquired tomorrow, can I keep working?" and added "Export buttons aren't boring features anymore."
Then the vendors arrived. ryan_morr quoted Holtz with "pov: you switched to cloud agents", which drew 41 likes, and followed with a migration link to Hoplite and a promo code for three months free. The account FlyaKiet replied to people asking for alternatives, one after another, saying Paradev imports Conductor workspaces; we counted at least 15 such replies between about 11:21 PM PT on November 4 and 2:13 AM PT on November 5. Orca, Tutti and AgentsRoom also posted their own pitches. None of these claims were tested by us.
The market read
A smaller group zoomed out. Sonath_AI wrote that "The coding agent startups are all being acquired this week," listing HumanLayer to Factory, Conductor to Meta and Pierre to OpenAI. We covered the last one in OpenAI to Acquire Pierre. EgeOnurGulec wrote that the dev tooling market "is consolidating pretty fast around really big model agnostic platforms like Devin and Factory or model providers like SpaceX and Meta."
On Hacker News, maherbeg called it "a great acquire for meta to super speed their coding agent rival", and sksum argued that the reasons companies buy startups, in order of real value, are talent, a fast-growing user base and speed against competitors. forthwall said the economics of these harnesses favor the large model providers, whose own usage plans will undercut small players. inshard asked whether Claude's desktop app already does all of this and more, and whether this layer is "super commoditized." For background on the tools in this category, see our explainer What Is a Coding Agent?.
What to take from it
Nobody we found disputed the deal's logic, and many replies congratulated the team. The complaints are about the short runway for Conductor Cloud users, the lack of an open source release, and the lack of a way to try the app now. The praise is specific: worktree setup, parallel agents, a combined pool of token limits, and a UI some users say they have not found elsewhere. Neither Conductor nor Meta has said what the new product will be, so the posts that look ahead are guesses.
The practical step is in Conductor's own FAQ: export your data or use "Transfer to local" before November 13. Posts like TbYoutuber's suggest the second lesson, which is to check a replacement tool's export options before moving a team onto it.
Counts and likes above are from the early morning of November 5 and will move.
