Photon Raises $4.5M After Its “App Funeral” Pushes Messaging Agents Into the Spotlight

Photon, the startup building AI agents that operate inside iMessage, WhatsApp and other messaging channels, has raised $4.5 million in seed funding as it pushes a future where consumers and businesses use agents instead of downloading more apps. The funding comes after the company says it signed up more than 40,000 developers and increased revenue tenfold in just four months, underscoring how quickly interest in messaging-based agent tools is growing.

The San Francisco company has become known for an unusual bit of product theater: it staged a literal funeral for mobile apps at a church on September 17, complete with speeches and a coffin filled with app icons. The event was part developer gathering, part publicity stunt, and part declaration of intent. Photon’s founders say the app era is far from over, but they believe the long-term center of gravity is moving toward agents that live inside the communication tools people already use every day.

That pitch has clearly resonated. Photon says its open-source technology still accounts for the overwhelming majority of its use, but the company has now converted enough developer attention into a hosted product, paid subscriptions and enterprise-ready features to attract new investors. The latest financing round was led by Gradient and A*, with participation from Vercel, HongShan, Z Fellows, Llama Ventures, Karman and several angel investors.

Why Photon is betting against the classic app model

Photon’s core argument is simple: people do not want to keep installing new standalone apps for every task. Instead, they want fast access to AI agents through channels they already open constantly, such as iMessage, WhatsApp, Telegram, SMS, RCS, email and voice.

Co-founder and chief executive Daniel Tian says the shift may take time, but he believes it is directionally unavoidable. In his view, the future is less about browsing app stores and more about invoking an agent inside a conversation. That could mean asking a messenger-based assistant to book travel, sort through customer support requests, send updates or coordinate with other automated systems on a user’s behalf.

The company is not alone in seeing potential in that model. Meta’s Muse, which has surged to the top of app store charts, suggests that consumer interest in agentic experiences is real. Still, Tian argues that even if a consumer product becomes dominant, there will remain a large market for infrastructure that lets developers deploy their own agent experiences across messaging platforms.

How the product works

Photon offers a toolkit designed to help developers build and manage agents that can communicate across multiple channels from a single stack. The platform includes a unified API, an extensible channel framework, a command-line interface and an observability suite.

In practical terms, that means a developer can create one agent and then extend it across messaging environments without rebuilding the entire system for each one. Photon says the platform can support experiences over iMessage, WhatsApp, Telegram, SMS, RCS, email and voice, giving teams a way to meet users where they already are.

The company’s managed service also adds reliability and compliance features that are harder to achieve with a local open-source build. Photon says the hosted version offers 99.95% uptime and is SOC 2 Type II and HIPAA-compliant, which makes it more attractive for customer-facing, regulated and healthcare-related use cases.

What fueled the startup’s rise?

Photon’s growth began with a problem many young builders recognize: making people discover and actually use a new product. Tian and co-founder Ryan Zhu, who is now CTO, spent years building consumer apps as students, often at hackathons. Each new release brought the same frustration — distribution was harder than development.

The turning point came when they realized that an agent embedded in iMessage could reach users through an app they already relied on every day. Tian had already experimented with a bot that could text his friends back in his own voice over iMessage. When the two founders released the technology on GitHub, thousands of developers started using it, revealing a broader demand for the underlying infrastructure.

That traction pushed both founders to commit fully. Tian left the University of Pennsylvania’s M&T program, while Zhu left high school and later became a visiting student at MIT’s Media Lab. Their early bet on messaging-native agents has since matured into a platform now used by developers building consumer and business tools.

How big is Photon’s developer base?

Photon says it has more than 40,000 developer sign-ups across its platform, with paying customers now using the hosted product. The company has not disclosed exact revenue, but it says revenue has grown tenfold over the last four months and messaging volume has increased fivefold in the past month.

The company also says customer churn is below 3%, a figure that suggests strong retention for a young infrastructure product. While many sign-ups likely come from experimentation, Photon appears to be converting at least some of that attention into recurring usage and paid deployment.

The free tier covers smaller use cases, but Photon says teams that exceed 10 users must move to one of the paid plans. That structure helps the company balance open-source adoption with a commercial path for teams building production systems.

Who is using Photon?

Photon says its customers include Corgi Insurance, social introductions app Boardy, Gen Z dating platform Ditto, business banking company Rho, finance assistant Fliptexts and AI email client Slashy. Those examples point to a broader pattern: the platform is finding early traction in fintech, consumer AI and other vertical applications where conversational interfaces can shorten the path from question to action.

The company’s partners and integrations also suggest it has become part of a wider ecosystem. Vercel uses Photon through products such as Eve and ChatSDK, while Nous Research’s Hermes agent uses Photon as its default iMessage layer. Photon also says it underpins Tencent’s QClaw and NanoClaw, and that it has integrations across LangChain, Mastra, Convex, Render, Railway and Telnyx.

Tian said the broader direction of the market is moving toward agents, even if the transition will take years, and described the long-term opportunity as a communication layer that connects people to the right automated tools when they need them.

Why investors are paying attention

Investors tend to back infrastructure that sits close to a major platform shift, and Photon is pitching itself as exactly that kind of company. If agents become a common interface for consumer and business tasks, the companies that make those agents easier to deploy, monitor and secure could become essential picks-and-shovels plays.

Photon’s appeal is not just the idea of agent-native messaging. It is the combination of open-source virality, measurable developer adoption and a managed product that can be sold to teams that want reliability and compliance. The startup is trying to show that its early popularity is not just developer curiosity, but the beginning of a sustainable platform business.

That is also why the funding round includes strategic names from the developer ecosystem. Vercel’s participation, in particular, links Photon to one of the better-known brands in modern application development and suggests the company’s messaging-agent thesis has support from adjacent infrastructure players.

What does the app funeral actually mean?

The “app funeral” was meant to be provocative, but it also functioned as a branding exercise for a larger argument: the software interface of the future may look less like rows of icons and more like conversational systems that act on behalf of users.

At the San Francisco event, Photon framed the gathering as a developer day as much as a symbolic send-off for mobile apps. The lineup featured speakers from Vercel, Stripe and OpenAI, reinforcing the idea that the company sees itself inside a larger ecosystem of tooling, payments and AI model providers rather than as a standalone chatbot maker.

The coffin full of app icons was a visual punchline, but the message was serious. Photon wants to be known as the infrastructure layer that makes it easy to build usable agents in the channels people already trust and check repeatedly throughout the day.

Why messaging apps matter so much

Messaging apps are valuable because they already command attention, identity and habit. Unlike a new app that must be downloaded, learned and remembered, a messaging thread is already part of the user’s daily flow. That lowers friction for discovery and ongoing engagement.

For developers, this matters because consumer software often fails not from a lack of usefulness, but from a lack of distribution. Photon’s founders came to this conclusion through their own experience building products that were technically interesting but difficult to get noticed. Messaging channels offered a shortcut to the attention problem.

There is also a practical advantage for agents that rely on asynchronous conversation. Messaging is well suited to tasks that do not require a full standalone interface, especially when an assistant can interpret a request, execute a workflow and then send a result back in the same thread.

What comes next for Photon?

Photon is now trying to move beyond novelty and prove that messaging-native agents can become a durable category. That means turning open-source popularity into a broader commercial business, expanding its hosted platform and supporting the kind of production-grade use cases that demand uptime, monitoring and compliance.

The company also appears to be looking ahead to a more networked agent economy. Tian says he thinks the future may involve agents finding other agents on behalf of users, creating an “agent-to-agent” communication layer that sits between people and the software systems capable of completing tasks.

In a travel scenario, for example, a person could speak with one assistant that then calls on a separate flight-booking agent or hotel-booking agent to complete parts of the request. Photon sees that as the start of an “A-to-A-to-P” model: agent to agent to person.

That idea may sound futuristic, but it fits the company’s broader strategy. Photon is not just selling a tool for chatting with bots. It is selling a way to make conversational software easier to build, deploy and connect across the channels where users already spend their time.

Timeline: Photon’s path from open source to seed funding

• Student projects: Tian and Zhu build consumer apps and repeatedly run into distribution problems.
• iMessage experiment: The founders test an agent that can operate inside Apple’s Messages app.
• GitHub release: They open-source the technology, attracting thousands of developers.
• Founder commitment: Tian leaves Penn’s M&T program and Zhu leaves high school, later spending time at MIT’s Media Lab.
• April 2026: Photon launches a hosted, subscription-based platform.
• September 17, 2026: The company stages its app funeral and developer day in San Francisco.
• October 1, 2026: Photon announces $4.5 million in seed funding.

Why this matters beyond one startup

Photon’s raise is a reminder that the AI industry is no longer just competing on model quality. It is also competing on interface design, distribution and the everyday places where users interact with software. Messaging apps may prove to be one of the most important battlegrounds for that shift.

If agents do become the preferred way to complete tasks, then the winners may be the companies that make those agents easy to deploy across channels, safe to operate in regulated settings and useful enough to replace some of the friction of app-based workflows. Photon wants to be one of those companies.

The startup’s future is still early and the “funeral” framing is intentionally theatrical. But the underlying business case is increasingly clear: build where users already are, reduce the cost of creating intelligent workflows, and make it simple for developers to meet demand before it fully materializes.

For now, Photon has one of the clearest bets in the emerging agent infrastructure race. It believes apps are not disappearing overnight, but it is wagering that more and more software interactions will happen inside conversations rather than outside them — and investors are willing to give that thesis a shot.