Photon Raises $4.5M Seed to Put AI Agents Inside Messaging Apps
Photon, an artificial intelligence startup whose tools let developers create agents that run inside iMessage, WhatsApp and similar services, has raised $4.5 million in seed funding. The company says developer sign-ups have passed 40,000 and that its revenue multiplied tenfold within four months.
To signal its conviction that agents will displace mobile applications, Photon staged what it described as an app funeral at a San Francisco church on September 17. The ceremony featured speeches, panels from Vercel, Stripe and OpenAI, and a coffin containing app icons. Co-founder and chief executive Daniel Tian conceded the event was partly in jest, yet argued the app era is nearing its end. "Honestly, it's going to take a while, but I think, directionally, it's inevitable," he said of the move from apps to agents.
Photon's premise is that consumers have grown weary of downloading new software and would prefer agents that operate through messaging services they already open daily, such as iMessage and WhatsApp. Meta's Muse, which has reached the top of the app stores, is testing that proposition. Even if Muse ends up dominating consumer AI, Tian sees an opening in catering to developers who want their products to function over messaging platforms.
To serve them, Photon provides a set of tools: a unified API, an extensible channel framework, a command-line interface and an observability suite. Used together, these allow developers to build and run agent experiences across iMessage, WhatsApp, Telegram, SMS or RCS, email, voice and additional channels.
The company's origin lies in the founders' student days, when they repeatedly built consumer apps, often at hackathons. Each launch ran into the same problem: helping users find the app. Building an agent that lived inside iMessage, they concluded, offered a route to users through an app already on their phones every day, Apple's Messages.
Tian had created a bot that answered his friends over iMessage in his voice. After he and co-founder Ryan Zhu, the chief technology officer, published the technology as open source on GitHub, thousands of developers took it up. The pair committed fully: Tian left UPenn's M&T program, while Zhu exited high school and later became a visiting student at MIT's Media Lab. The open-source edition still accounts for 98% of Photon's usage.
In April, the founders introduced a managed platform and started charging through subscription tiers. The hosted product advertises 99.95% uptime, well above what a hobby project running the open-source version on a local Mac would achieve. It is also SOC 2 Type II and HIPAA-compliant, which creates room for healthcare agent applications.
More than 40,000 developers have signed up for the platform, which now has paying customers. Usage past 10 users requires leaving the free tier for one of three paid plans. Photon declined to disclose revenue, but alongside the 10x growth it reported churn below 3% and a fivefold rise in messaging volume over the past month alone.
Among the customers using Photon are Corgi Insurance; Boardy, a social introductions app; Ditto, a Gen Z dating platform; Rho, a business banking platform; Fliptexts, a finance AI assistant; and Slashy, an AI email client, along with others building agents for fintech, consumer AI and other verticals.
On the technology side, Photon's integration partners include Vercel (Eve and ChatSDK) and Nous Research, whose Hermes agent relies on Photon as its default iMessage layer. Photon also provides the layer beneath Tencent's QClaw and NanoClaw. The company has additionally recorded framework and infrastructure integrations with LangChain, Mastra, Convex, Render, Railway and Telnyx, among others.
Vercel is likewise among Photon's new investors. Gradient and A* co-led the seed round, with Vercel, HongShan, Z Fellows, Llama Ventures, Karman and other angel investors taking part.
Tian estimates that the technology, counting both the hosted and open-source versions, reaches millions of end users.
Although Photon currently concentrates on agents interacting with people, Tian sees room to grow into a market where agents connect with one another. "I think, in the future, agents will be able to find you the agent that can do the right things — so it's going to be the A-to-A communication layer," he said. As an example, he described a person chatting with an AI assistant about travel plans, where the assistant could call in other agents as required, such as one for booking flights or hotels. "It's been pretty powerful to observe the whole A-to-A-to-P [agent-to-agent-to-person] space," he added.
to like, bookmark, and comment.
Comments
Join the conversation:
No comments yet. Be the first to share your thoughts.