Conductor was a two-year-old Mac app that let developers run several AI coding agents at once, each working on its own copy of a codebase, so you could compare what Claude Code, Codex, Cursor and OpenCode each produced and merge the best result. On November 4, Meta announced it was buying the company. Financial terms were not disclosed. What was disclosed, within days, was a shutdown calendar: Conductor Cloud goes read only on November 13 and closes for good on November 20, with user data deletion starting right after.
What actually got bought, and what did not
Conductor was founded in 2024 by CEO Charlie Holtz and CTO Jackson de Campos. The company raised a seed round and then a $22 million Series A in March, led by Matrix and Spark, with Y Combinator and the founders of Notion and Linear also putting money in. Its core cloud product launched in July 2025, so it had been generally available for a little over a year when Meta came calling.
Meta is not keeping the business running as a business. It is keeping the people and the technology. Holtz said he is looking forward to working with Alexandr Wang, Meta's chief AI officer, and Nat Friedman, who leads product for Meta Superintelligence Labs. That is the clearest signal of what this deal actually is: Meta did not buy a product it plans to sell to outside customers, it bought a team that already solved a problem Meta's own coding tools still need solved.
The shutdown timeline, in plain terms
For anyone who was actually paying for Conductor, the practical details matter more than the strategic ones. Here is what the company laid out:
- November 13: Conductor Cloud becomes read only, so you can still see old chats and export your data, but you cannot start new cloud sessions
- November 20: The cloud platform shuts down completely and account data deletion begins
- Early 2027: The local Mac app stops getting new features and meaningful bug fixes, though it keeps working for now
Pro subscribers are getting refunded, and free users can no longer upgrade to a paid tier, since there is nothing left to upgrade into. People who want to keep using the tool at all are being pointed toward the standalone local app, which runs on your own machine instead of Conductor's cloud.
Why a small coding app is worth buying at all
Meta launched its own coding agent, Muse Code, in August, and has been pushing its broader Muse agent platform hard since. Running multiple AI coding agents in parallel on isolated copies of a codebase, then reviewing and merging the results, sounds like a small feature. It is actually one of the harder interface problems in AI coding right now, because agents frequently touch the same files, make conflicting edits, or need a human to referee between two reasonable-looking solutions. Conductor had already shipped a working answer to that problem for Mac, iPhone and cloud use, across four different agent providers.
That makes the deal look less like Meta wanting a product and more like Meta wanting a working solution to a specific technical headache, built by a small team, fast. It joins a run of similar Meta moves this year, including bringing on the founders of AI safety startup Virtue AI over the summer. The pattern is consistent: identify a small team that already cracked a narrow, hard problem relevant to Meta's AI push, then absorb it.
The part that gets skipped in the headlines
Most coverage of this deal led with Meta's AI ambitions, which is the investor-relevant angle. The less covered angle is what it is like to be a customer in the middle of it. Conductor had real paying subscribers who built workflows around a tool that, two weeks after the acquisition was announced, told them to export their data before it gets deleted. The company is handling the wind-down reasonably, with refunds and a migration path to a local-only version, but the compressed timeline is the part that rarely makes it into funding-round writeups.
This is not unique to Conductor. It is close to the default outcome when a venture-backed tool gets bought primarily for its team rather than its customer base. The product usually survives in some diminished form for a while, then quietly stops being maintained.
The takeaway
The lesson here is not about Meta specifically, it is about what "undisclosed acquisition" tends to mean in practice when you are an actual user of the thing being acquired. A funding round or an acquisition headline almost never tells you whether the product you depend on is being kept, folded into something else, or wound down on a two-week clock. If you are relying on a smaller AI tool for real work, it is worth asking which of those three outcomes you could live with, because the deal that makes the founders and investors happy is not always the one that keeps your subscription running.