Perplexity is throwing its first-ever developer conference this March, and the name alone—Ask—pretty much sums up the mood: a company that’s been quietly building, now ready to show its cards in front of the people who can actually do something with them. It’s not a sprawling, multi-day spectacle like Google I/O or WWDC; it’s one morning in San Francisco, invitation-only, with roughly 100 builders in the room and the promise that the hood is finally coming up on what Perplexity has really been working on.

The basics are deliberately restrained: Ask 2026 is scheduled for March 11, 2026, in San Francisco, running from 10 am to 2 pm, with the specific venue revealed only after registration is confirmed. The event page makes it clear this is not a mass-market keynote; it’s “by invitation-only,” framed as a closed-door, front-row seat for builders and leaders who move first and build differently. There’s also a waitlist and an open call for standout developers who may not yet be on Perplexity’s radar, signaling that they’re trying to balance exclusivity with genuine community discovery.

What’s interesting is the timing. Ask lands at the end of a sprint, where Perplexity has been shipping at a pace that looks less like a search startup and more like an emerging AI platform operator. Perplexity recently rolled out a dedicated Search API, giving developers access to the same global-scale index that powers its consumer answer engine—hundreds of billions of webpages wired into an AI-optimized search stack, complete with an SDK, an open-source evaluation framework, and a new API platform to house both Search and Sonar. In parallel, Perplexity has aggressively positioned its APIs into distribution deals: its models now power features in hundreds of millions of Samsung devices and across multiple “Magnificent Seven” tech giants, a sign that the company isn’t just courting indie hackers but also embedding itself deep inside the existing tech stack.

Then there’s Perplexity Computer, the company’s latest swing at the agentic future. It’s pitched as a “computer user agent” that can independently execute complex workflows by orchestrating 19 different AI models, spinning up sub-agents as needed to handle specific parts of a task. Instead of focusing on a single frontier model, Perplexity is leaning into a multi-model philosophy, arguing that the real advantage is in how these models are routed, combined, and evaluated for cost, speed, and accuracy depending on what the user is trying to do—whether that’s deep research, software engineering, or data-heavy analysis. That agentic layer is only available on the top-tier Perplexity Max plan for now, but it’s a clear signal about where the company thinks the interface between humans and AI is headed.

Ask, in that context, looks less like a marketing event and more like a strategy milestone. Perplexity started out in the public imagination as an “AI search” product—a faster, more conversational way to look things up, with good citations and a clean interface. Over the last year, though, the story has shifted. With the Search API, Sonar models, a unified API platform, and now an agentic “Computer,” Perplexity increasingly resembles a developer platform and, in the more ambitious framing, a potential operating layer for AI-powered applications. A developer conference is the next logical move: it formalizes the ecosystem, gives partners and early adopters a place to compare notes, and, crucially, offers Perplexity a stage to debut whatever is coming next.

The official description of Ask is deliberately teaser-like: “We’ve been quiet. Building.” Founders will take the stage, new products will launch, and developers already building on Perplexity will share what’s actually working in production. That last part matters. Perplexity’s pitch to devs has always leaned on practicality: real-time web access, source citations, and APIs that slot into existing OpenAI-style client libraries, rather than a totally bespoke integration path. Showcasing teams who are already shipping on top of Perplexity—whether it’s research tools, data products, or agents wired into business workflows—turns the conference into a living proof-of-concept gallery.

There’s also an ecosystem and governance subtext here that’s hard to ignore. As more AI tools move from answering questions to actually taking actions—triggering workflows, touching sensitive data, making decisions—the questions of auditing, control, and reliability stop being theoretical. Some in the AI community have already framed Perplexity’s evolution as a platform play where compute, distribution, and ecosystem form the moat, but long-term defensibility also depends on how well these systems can be governed and monitored at scale. A developer conference is a natural place to talk, even informally, about how Perplexity thinks about safe execution, boundary control, and the operational side of letting agents loose inside enterprise and consumer environments.

For developers, the appeal of Ask is pretty straightforward. You get a concentrated, half-day download directly from Perplexity’s founders and product builders, plus a room full of peers who are already pushing the API in different directions—from retrieval-heavy apps to research dashboards to autonomous assistants. You also get a first look at whatever’s next: more models, new agent capabilities, deeper integrations, or entirely new surfaces like Perplexity Comet, a browser experience is headed to iOS. In a world where every AI company is racing to define what “platform” actually means, being in that room on March 11 is a way to place an early bet, or at least to see firsthand how serious Perplexity is about being more than a really good answer engine.

Ask 2026 isn’t trying to be a festival. It’s a four-hour, closed-door briefing for people who see Perplexity not just as a tab they open in their browser, but as infrastructure they want to build on. If Perplexity’s recent moves are any indication—the Search API, the API Platform, Perplexity Computer, and its widening distribution footprint—this first developer conference is less about celebration and more about alignment: putting developers, partners, and the company itself in the same room and asking, quite literally, what they want to build next.