NEWS

Meta Buys Conductor and Begins Shutting Down the Multi-Agent Coding Tool

Meta Buys Conductor and Begins Shutting Down the Multi-Agent Coding Tool
Image: Redação iMasters
News
Reading
5 min
In short

Meta confirmed the acquisition of the startup, which let users run multiple AI agents in parallel on the same project. The current app is being shut down in stages through early 2027, and users already have a deadline to migrate their data.

Meta confirmed the acquisition of Conductor, a startup that let software engineers run multiple AI agents in parallel workspaces to advance the same project. The acquisition was confirmed by Meta to TechCrunch and by Charlie Holtz, founder and CEO of Conductor, in a post on the company's blog. The deal's value was not disclosed, and the startup will be folded into Superintelligence Labs (MSL), Meta's AI division.

For those already using the tool, the news matters less for the valuation and more for the timeline: the current product is being shut down in stages, with deadlines starting this week. For those who don't use it, it matters for what it signals: Meta is buying up developer tooling at a pace that already adds up to at least three AI acquisitions in 12 months.

What Conductor Did

Founded in 2024, Conductor launched its product in July 2025. The idea was to let a team of agents work in a coordinated way on different parts of the same project, triggered from the cloud, a local machine, or a phone. The company didn't build the agents itself: they were supplied by labs such as Anthropic and OpenAI, and Conductor handled the orchestration and the interface between them.

It's a different model from the single in-editor copilot, like GitHub Copilot or Claude Code: Conductor sold the idea of multiple agents running in parallel, each one on a different front of the code, supervised by whoever orchestrates them. Holtz described the moment of launch as counterintuitive:

Writing most of your code with agents, let alone with a team of them, seemed crazy back then.

Charlie Holtz, founder and CEO of Conductor

Shutdown Timeline

Meta has already begun gradually shutting down the service. The dates, according to Conductor's announcement:

  • November 4, 2026: new sign-ups and app downloads are blocked; free-plan users can no longer upgrade to the paid plan; pro-plan subscribers keep pro features at no cost.
  • November 13, 2026: Conductor Cloud becomes read-only. You can still view old chats and export data, but not send new messages. Active subscribers receive a refund.
  • November 20, 2026: the cloud service is shut down and the company begins deleting stored data.
  • Early 2027: the local version, which continues receiving new models and critical bug fixes until then, is also shut down.

Anyone relying on Conductor in production has less than three weeks of read-only access in the cloud and just over two months until the local version reaches its end.

How to Migrate Before the Deadline

Conductor recommends two paths for those who want to preserve their work:

  1. Move a workspace from the cloud to your computer: right-click the workspace and choose "Transfer to local".
  2. Download all account data: go to Settings, then Account, then Export data.

The company asks that the export be done before November 20, when the cloud service goes offline and data deletion begins. According to Conductor, the company has never sold or used user data to train models, and won't start now: all content will be deleted once the app is permanently shut down in early 2027.

For teams running automated pipelines on top of the Conductor API, the practical message is straightforward: treat migration to the local version, or to another orchestration tool, as a short-term item, not a 2027 problem. The local version lasts longer, but it has an expiration date.

What Comes Next

Holtz was vague about what will replace the current product, but promised news for early next year:

We have ambitious, kind of crazy plans for the next interface. If you liked Conductor, you're going to love what's coming.

Charlie Holtz, founder and CEO of Conductor

In a post on X, Holtz said he expects to work directly with Alexandr Wang, Meta's Chief AI Officer, and Nat Friedman, head of product at MSL. He praised the division's pace:

They're moving at an overwhelming pace, and I'm convinced MSL has the highest density of talent of any lab.

Charlie Holtz, founder and CEO of Conductor

This suggests that whatever replaces Conductor won't be a simple continuation of the current product, but something rebuilt within Meta's own AI infrastructure, likely integrated with Muse Code.

Where This Fits Into Meta's Strategy

The Conductor acquisition isn't an isolated move. In August this year, Meta launched Muse Code, its own agent aimed at large codebases, and the company reported that Muse agents already have more than 10 million users.

In September, Meta bought Stilla.ai, a Swedish startup founded in 2024 in Stockholm that develops AI agents for businesses and had been funded by General Catalyst and angel investors linked to ElevenLabs, Lovable, Shopify, Legora, Synthesia, Sana Labs, and Hugging Face; the acquisition aims to accelerate Meta Business Agent. Over the past 12 months, the company also bought Moltbook, an agent platform that had gained its own notoriety.

The pattern is clear: Meta, which had been falling behind in the race for productivity-focused AI services, is buying entire teams instead of building each piece from scratch. For the Brazilian developer, this means one more player competing for budget and attention on the same shelf where GitHub Copilot, Claude Code, and Cursor already sit, with the difference that Meta is coming in through the multi-agent orchestration route, not the in-editor copilot.

What Remains Open

The announcement doesn't detail whether Meta's new product will keep offering third-party agents, as Anthropic and OpenAI did within Conductor, or whether everything will be pushed into Muse Code. There's also no pricing or exact launch date beyond "early next year." For anyone deciding on a stack today, the practical signal is this: teams that bet on Conductor as the central piece of their workflow need a plan B now, and anyone evaluating multi-agent orchestration tools gains one more big competitor on the radar, still without a public product to test.

Translated from the Brazilian Portuguese original · Read the original